diff --git a/ansible/playbooks/paas/coredns.yml b/ansible/playbooks/paas/coredns.yml
index 284f0a85..85dff61f 100644
--- a/ansible/playbooks/paas/coredns.yml
+++ b/ansible/playbooks/paas/coredns.yml
@@ -5,5 +5,4 @@
   gather_facts: true
   become: true
   roles:
-    - golang
     - coredns
diff --git a/ansible/playbooks/paas/main.yml b/ansible/playbooks/paas/main.yml
index 85c0b7f3..b1c0b2e4 100644
--- a/ansible/playbooks/paas/main.yml
+++ b/ansible/playbooks/paas/main.yml
@@ -5,6 +5,12 @@
   gather_facts: true
   become: true
   pre_tasks:
+
+    - name: Set fqdn hostname
+      ansible.builtin.hostname:
+        name: "{{ inventory_hostname }}"
+        use: systemd
+
     - name: Create ansible facts.d directory
       become: true
       ansible.builtin.file:
@@ -14,20 +20,6 @@
         group: "root"
         mode: '0755'
 
-    - name: Get ipinfo.io
-      ansible.builtin.uri:
-        url: https://ipinfo.io
-        http_agent: curl/7.81.0
-      register: register_uri
-      check_mode: false
-
-    - name: Set ipinfo local_fact
-      ansible.builtin.copy:
-        content: |
-          {{ register_uri.json | to_nice_json }}
-        dest: /etc/ansible/facts.d/ipinfo.fact
-        mode: '0644'
-
     - name: Install mandatories packages
       ansible.builtin.apt:
         pkg:
@@ -42,5 +34,17 @@
       until: apt_status is success
       delay: 6
       retries: 10
+
   roles:
     - unattended-upgrades
+
+- name: Configure systemd resolved
+  ansible.builtin.import_playbook: systemd-resolved.yml
+- name: Configure docker
+  ansible.builtin.import_playbook: docker.yml
+- name: Configure nomad
+  ansible.builtin.import_playbook: nomad.yml
+- name: Configure coredns
+  ansible.builtin.import_playbook: coredns.yml
+- name: Configure metrology
+  ansible.builtin.import_playbook: metrology.yml
diff --git a/ansible/playbooks/paas/metrology.yml b/ansible/playbooks/paas/metrology.yml
index 2f334268..56718234 100644
--- a/ansible/playbooks/paas/metrology.yml
+++ b/ansible/playbooks/paas/metrology.yml
@@ -4,16 +4,51 @@
   hosts: "{{ hosts_limit | default('infrastructure') }}"
   gather_facts: true
   become: true
-  roles:
-    - prometheus
-    - promtail
-    - phpfpm_exporter
-    - node_exporter
-    - mysqld_exporter
-    - systemd_exporter
-    - mongodb_exporter
-    - blackbox_exporter
-    - nginx_exporter
-    - scan_exporter
-    - dns_exporter
-    - script_exporter
+  vars_prompt:
+    - name: project
+      prompt: project name
+      private: false
+  tasks:
+    - name: End the play for hosts that are not in admins group
+      ansible.builtin.meta: end_host
+      when: fact_instance.location != 'admins'
+
+    - name: Install prometheus
+      ansible.builtin.import_role:
+        name: prometheus
+
+- name: Install exporters
+  any_errors_fatal: true
+  hosts: "{{ hosts_limit | default('infrastructure') }}"
+  gather_facts: true
+  become: true
+  tasks:
+    - name: Create prometheus group
+      ansible.builtin.group:
+        name: prometheus
+        system: true
+
+    - name: Create prometheus user
+      ansible.builtin.user:
+        name: prometheus
+        create_home: false
+        system: true
+
+    - name: Install exporters
+      ansible.builtin.include_role:
+        name: "{{ exporter }}"
+      loop:
+        - promtail
+        - phpfpm_exporter
+        - node_exporter
+        - mysqld_exporter
+        - systemd_exporter
+        - mongodb_exporter
+        - blackbox_exporter
+        - nginx_exporter
+        - scan_exporter
+        - dns_exporter
+        - script_exporter
+        - nvidia_gpu_exporter
+      loop_control:
+        loop_var: exporter
diff --git a/ansible/playbooks/paas/nomad-juicefs.yml b/ansible/playbooks/paas/nomad-juicefs.yml
new file mode 100644
index 00000000..f4de3129
--- /dev/null
+++ b/ansible/playbooks/paas/nomad-juicefs.yml
@@ -0,0 +1,11 @@
+---
+- name: Install nomad juicefs CSI driver
+  any_errors_fatal: true
+  hosts: "{{ hosts_limit | default('infrastructure') }}"
+  gather_facts: true
+  become: true
+  tasks:
+    - name: Install nomad juicefs CSI driver
+      ansible.builtin.import_role:
+        name: nomad
+        tasks_from: 10_juicefs
diff --git a/ansible/playbooks/paas/nvidia.yml b/ansible/playbooks/paas/nvidia.yml
new file mode 100644
index 00000000..d8f2314d
--- /dev/null
+++ b/ansible/playbooks/paas/nvidia.yml
@@ -0,0 +1,142 @@
+---
+- name: Install nomad nvidia plugin
+  any_errors_fatal: true
+  hosts: "{{ hosts_limit | default('infrastructure') }}"
+  gather_facts: true
+  become: true
+  vars:
+    build_work_dir: /tmp
+    upstream_file_url: https://github.com/hashicorp/nomad-device-nvidia.git
+    nvidia_container_toolkit_version: "1.17.8-1"
+    nvidia_gpg_key_url: "https://nvidia.github.io/libnvidia-container/gpgkey"
+    nvidia_repo_list_url: "https://nvidia.github.io/libnvidia-container/stable/deb/nvidia-container-toolkit.list"
+    nvidia_keyring_path: "/usr/share/keyrings/nvidia-container-toolkit-keyring.gpg"
+    nvidia_list_path: "/etc/apt/sources.list.d/nvidia-container-toolkit.list"
+
+  roles:
+    - golang
+
+  pre_tasks:
+
+    - name: Créer le répertoire du keyring s'il n'existe pas
+      ansible.builtin.file:
+        path: "{{ nvidia_keyring_path | dirname }}"
+        state: directory
+        mode: "0755"
+
+    - name: Télécharger et enregistrer la clé GPG NVIDIA
+      ansible.builtin.get_url:
+        url: "{{ nvidia_gpg_key_url }}"
+        dest: /tmp/nvidia-container-toolkit.gpg
+        mode: "0644"
+
+    - name: Convertir la clé GPG en format keyring
+      ansible.builtin.command:
+        cmd: "gpg --dearmor -o {{ nvidia_keyring_path }} /tmp/nvidia-container-toolkit.gpg"
+        creates: "{{ nvidia_keyring_path }}"
+
+    - name: Télécharger le fichier de dépôt NVIDIA et ajouter le signed-by
+      ansible.builtin.shell: |
+        curl -s -L {{ nvidia_repo_list_url }} | \
+        sed 's#deb https://#deb [signed-by={{ nvidia_keyring_path }}] https://#g' > {{ nvidia_list_path }}
+      args:
+        creates: "{{ nvidia_list_path }}"
+
+    - name: Activer la section experimental (décommenter)
+      ansible.builtin.replace:
+        path: "{{ nvidia_list_path }}"
+        regexp: '^#(.*experimental.*)$'
+        replace: '\1'
+
+    - name: Mettre à jour la liste des paquets
+      ansible.builtin.apt:
+        update_cache: true
+
+    - name: Installer les paquets NVIDIA Container Toolkit
+      ansible.builtin.apt:
+        name:
+          - "nvidia-container-toolkit={{ nvidia_container_toolkit_version }}"
+          - "nvidia-container-toolkit-base={{ nvidia_container_toolkit_version }}"
+          - "libnvidia-container-tools={{ nvidia_container_toolkit_version }}"
+          - "libnvidia-container1={{ nvidia_container_toolkit_version }}"
+        state: present
+
+  tasks:
+    - name: Install dependencies
+      ansible.builtin.apt:
+        pkg:
+          - nvidia-utils-580
+          - nvidia-driver-580
+          - nvidia-container-runtime
+          - nomad-device-nvidia
+        state: present
+        install_recommends: true
+        update_cache: true
+      register: apt_status
+      until: apt_status is success
+      delay: 6
+      retries: 10
+
+    - name: Nomad-nvidia-plugin | Git checkout
+      ansible.builtin.git:
+        repo: https://github.com/hashicorp/nomad-device-nvidia.git
+        dest: "{{ build_work_dir }}/nomad-device-nvidia"
+        version: main
+        force: true
+
+    - name: Nomad-nvidia-plugin | Build binary
+      ansible.builtin.command:
+        cmd: make compile
+        chdir: "{{ build_work_dir }}/nomad-device-nvidia"
+      environment:
+        PATH: "/usr/local/go/bin:{{ ansible_env.PATH }}"
+      register: my_output
+      changed_when: my_output.rc != 0
+
+    - name: Create nomad plugin directory
+      ansible.builtin.file:
+        path: /opt/nomad/plugins
+        state: directory
+        owner: root
+        group: root
+        mode: "0755"
+
+    - name: Nomad-nvidia-plugin | Copy binary
+      ansible.builtin.copy:
+        src: /tmp/nomad-plugins/nomad-device-nvidia
+        dest: /opt/nomad/plugins/nomad-device-nvidia
+        owner: root
+        group: root
+        mode: '0755'
+        remote_src: true
+
+
+    - name: Copy using inline content
+      ansible.builtin.copy:
+        content: |
+          plugin "nomad-device-nvidia" {
+            config {
+              enabled            = true
+              fingerprint_period = "5s"
+            }
+          }
+        dest: /etc/nomad.d/nvidia.hcl
+        owner: root
+        group: root
+        mode: '0644'
+
+    # https://docs.nvidia.com/datacenter/cloud-native/container-toolkit/latest/install-guide.html
+    - name: Nomad-nvidia-plugin | Test nvidia support
+      ansible.builtin.command: nvidia-ctk runtime configure --runtime=docker
+
+    - name: Nomad-nvidia-plugin | Restart docker
+      ansible.builtin.command: systemctl restart docker
+
+    # https://docs.nvidia.com/datacenter/cloud-native/container-toolkit/latest/sample-workload.html
+    - name: Nomad-nvidia-plugin | Test nvidia support
+      ansible.builtin.command: docker run --rm --runtime=nvidia --gpus all ubuntu nvidia-smi
+      register: docker_run
+
+    - name: Nomad-nvidia-plugin | Debug
+      ansible.builtin.debug:
+        msg: "{{ docker_run }}"
diff --git a/ansible/playbooks/paas/partition-data.yml b/ansible/playbooks/paas/partition-data.yml
new file mode 100644
index 00000000..eff0d7cd
--- /dev/null
+++ b/ansible/playbooks/paas/partition-data.yml
@@ -0,0 +1,32 @@
+---
+- name: Configure sdb partition
+  any_errors_fatal: true
+  hosts: "{{ hosts_limit | default('infrastructure') }}"
+  gather_facts: true
+  become: true
+  tasks:
+    - name: Create default directory
+      ansible.builtin.file:
+        path: /data
+        state: directory
+        owner: root
+        group: root
+        mode: '0755'
+
+    - name: Create a new primary partition /dev/sdb1
+      community.general.parted:
+        device: /dev/sdb
+        number: 1
+        state: present
+
+    - name: Create a ext4 filesystem on /dev/sdb1
+      community.general.filesystem:
+        fstype: ext4
+        dev: /dev/sdb1
+
+    - name: Mount up device
+      ansible.posix.mount:
+        path: /data
+        src: /dev/sdb1
+        fstype: ext4
+        state: present
diff --git a/ansible/playbooks/paas/roles/ansible-docker/defaults/main.yml b/ansible/playbooks/paas/roles/ansible-docker/defaults/main.yml
index 53e0e926..5123c51c 100644
--- a/ansible/playbooks/paas/roles/ansible-docker/defaults/main.yml
+++ b/ansible/playbooks/paas/roles/ansible-docker/defaults/main.yml
@@ -13,7 +13,9 @@ docker_private_registry_state: false
 docker_private_registry_url: ""
 docker_private_registry_username: ""
 docker_private_registry_password: ""
-docker_private_registry_config: /etc/docker/config.json
+docker_private_registry_config:
+  - /etc/docker/config.json
+  - /root/.docker/config.json
 
 # DNS
 docker_dns_configuration: true
diff --git a/ansible/playbooks/paas/roles/ansible-docker/tasks/install.yml b/ansible/playbooks/paas/roles/ansible-docker/tasks/install.yml
index 01bb2add..dd13b02f 100644
--- a/ansible/playbooks/paas/roles/ansible-docker/tasks/install.yml
+++ b/ansible/playbooks/paas/roles/ansible-docker/tasks/install.yml
@@ -21,14 +21,7 @@
     url: "https://download.docker.com/linux/{{ ansible_distribution | lower }}/gpg"
     dest: /etc/apt/keyrings/docker.asc
 
-- name: Add Docker repository on ubuntu < 24.04
-  ansible.builtin.apt_repository:
-    repo: "deb [arch={{ upstream_default_arch }} signed-by=/etc/apt/keyrings/docker.asc] https://download.docker.com/linux/ubuntu {{ ansible_distribution_release }} stable"
-    state: present
-    filename: docker
-  when: ansible_distribution_version is version('24.04', '<')
-
-- name: Add Docker repository on ubuntu >= 24.04
+- name: Add Docker repository on ubuntu
   ansible.builtin.copy:
     content: |
       Components: stable
@@ -42,7 +35,6 @@
     owner: root
     group: root
     mode: '0644'
-  when: ansible_distribution_version is version('24.04', '>=')
 
 - name: Install Docker
   ansible.builtin.apt:
@@ -68,14 +60,24 @@
     append: true
   notify: Docker_restart
 
+- name: Create home docker directory
+  ansible.builtin.file:
+    path: "{{ item }}"
+    recurse: true
+    state: directory
+    mode: '0755'
+  loop:
+    - /root/.docker
+    
 - name: Copy config.json
   ansible.builtin.template:
     src: config.json.j2
-    dest: "{{ docker_private_registry_config }}"
+    dest: "{{ item }}"
     owner: root
     group: root
     mode: '0600'
   when: docker_private_registry_state
+  loop: "{{ docker_private_registry_config }}"
   notify: Docker_restart
 
 - name: Copy daemon.json for DNS resolution
diff --git a/ansible/playbooks/paas/roles/coredns/README.md b/ansible/playbooks/paas/roles/coredns/README.md
index e03dc9f4..55beefb1 100644
--- a/ansible/playbooks/paas/roles/coredns/README.md
+++ b/ansible/playbooks/paas/roles/coredns/README.md
@@ -1,3 +1,7 @@
 # Role: `coredns`
 
 ## How to use this Ansible role?
+
+### nomad cluster mode
+
+nomad_primary_master_node: Set a primary nomad master node to get nomad_management_token
\ No newline at end of file
diff --git a/ansible/playbooks/paas/roles/coredns/tasks/build.yml b/ansible/playbooks/paas/roles/coredns/tasks/build.yml
index 86cc7946..fc9b0c1c 100644
--- a/ansible/playbooks/paas/roles/coredns/tasks/build.yml
+++ b/ansible/playbooks/paas/roles/coredns/tasks/build.yml
@@ -14,37 +14,18 @@
     url: "{{ upstream_file_url }}"
     dest: "{{ build_work_dir }}/download/"
     mode: '0644'
-    force: no
   register: download_result
 
-- name: Coredns | Git checkout
-  ansible.builtin.git:
-    repo: https://github.com/coredns/coredns
-    dest: "{{ build_work_dir }}/download/coredns"
-    version: master
-    force: true
-
 - name: Coredns | Unarchive GitHub release
   ansible.builtin.unarchive:
-    src: "{{ build_work_dir }}/download/{{ image.upstream.repo }}-{{ upstream_file_name }}"
+    src: "{{ build_work_dir }}/download/{{ upstream_file_name }}"
     dest: "{{ build_work_dir }}/download"
     remote_src: true
   when: download_result.changed
 
-- name: Coredns | Build binary
-  ansible.builtin.shell:
-    cmd: "{{ item }}"
-    chdir: "{{ build_work_dir }}/download/coredns"
-  environment:
-    PATH: "/usr/local/go/bin:{{ ansible_env.PATH }}"
-  loop:
-    - echo "nomad:github.com/ituoga/coredns-nomad" >> plugin.cfg
-    - go mod edit -replace github.com/ituoga/coredns-nomad={{ build_work_dir }}/download/coredns-nomad-{{ latest_version }}
-    - make gen coredns
-
 - name: Coredns | Copy binary
   ansible.builtin.copy:
-    src: "{{ build_work_dir }}/download/coredns/{{ image.upstream.binary }}"
+    src: "{{ build_work_dir }}/download/coredns"
     dest: /usr/local/bin/coredns
     owner: root
     group: root
diff --git a/ansible/playbooks/paas/roles/coredns/tasks/main.yml b/ansible/playbooks/paas/roles/coredns/tasks/main.yml
index a7b94816..8f5799e0 100644
--- a/ansible/playbooks/paas/roles/coredns/tasks/main.yml
+++ b/ansible/playbooks/paas/roles/coredns/tasks/main.yml
@@ -3,13 +3,25 @@
   ansible.builtin.include_vars: upstream.yml
 
 - name: Coredns | Get binary
-  include_tasks: build.yml
+  ansible.builtin.include_tasks: build.yml
   when: ansible_local[image.name] is not defined or ansible_local[image.name] != latest_version
 
+- name: Coredns | Create group
+  ansible.builtin.group:
+    name: coredns
+    system: true
+
+- name: Coredns | Create user
+  ansible.builtin.user:
+    name: coredns
+    create_home: false
+    system: true
+
 - name: Coredns | Create custom directories
   ansible.builtin.file:
     dest: "{{ item.dest }}"
     state: directory
+    mode: '0755'
     owner: "{{ item.owner | default('root') }}"
     group: "{{ item.group | default('root') }}"
   loop:
@@ -19,7 +31,7 @@
   ansible.builtin.template:
     src: "{{ item.src }}"
     dest: "{{ item.dest }}"
-    mode: 0600
+    mode: '0640'
     owner: coredns
     group: coredns
   loop:
@@ -46,5 +58,5 @@
       Cache=no
       DNSStubListenerExtra=172.17.0.1:53
     dest: /etc/systemd/resolved.conf.d/coredns.conf
-    mode: 0644
+    mode: '0644'
   notify: Restart systemd-resolved
diff --git a/ansible/playbooks/paas/roles/coredns/templates/Corefile.j2 b/ansible/playbooks/paas/roles/coredns/templates/Corefile.j2
index 491b7723..5cf6dbf6 100644
--- a/ansible/playbooks/paas/roles/coredns/templates/Corefile.j2
+++ b/ansible/playbooks/paas/roles/coredns/templates/Corefile.j2
@@ -4,9 +4,8 @@ service.nomad.:1053 {
     #debug
     #log
     nomad {
-        zone service.nomad
-        address https://127.0.0.1:4646
-        token {{ lookup('simple-stack-ui', type='secret', key=inventory_hostname, subkey='nomad_management_token', missing='error') }}
+        address https://{{ hostvars[nomad_primary_master_node | default(inventory_hostname)]['ansible_ens3']['ipv4']['address'] }}:4646
+        token {{ lookup('simple-stack-ui', type='secret', key=nomad_primary_master_node | default(inventory_hostname), subkey='nomad_management_token', missing='error') }}
         ttl 10
     }
     prometheus 127.0.0.1:9153
diff --git a/ansible/playbooks/paas/roles/coredns/templates/postinst.j2 b/ansible/playbooks/paas/roles/coredns/templates/postinst.j2
deleted file mode 100644
index 6284ae86..00000000
--- a/ansible/playbooks/paas/roles/coredns/templates/postinst.j2
+++ /dev/null
@@ -1,27 +0,0 @@
-#!/bin/bash
-
-set -e
-
-SERVICE_NAME="{{ image.name }}"
-USER_NAME="coredns"
-GROUP_NAME="coredns"
-
-if ! getent group "$GROUP_NAME" >/dev/null; then
-  echo "Creating group $GROUP_NAME..."
-  groupadd --system $GROUP_NAME
-fi
-
-if ! id -u "$USER_NAME" >/dev/null 2>&1; then
-  echo "Creating user $USER_NAME..."
-  useradd --system --gid $GROUP_NAME --shell /usr/sbin/nologin --no-create-home $USER_NAME
-fi
-
-
-if command -v systemctl >/dev/null 2>&1; then
-  echo "Reloading systemd configuration..."
-  systemctl daemon-reload
-
-  echo "Enabling and starting $SERVICE_NAME service..."
-  systemctl enable "$SERVICE_NAME"
-  systemctl start "$SERVICE_NAME"
-fi
diff --git a/ansible/playbooks/paas/roles/coredns/templates/prerm.j2 b/ansible/playbooks/paas/roles/coredns/templates/prerm.j2
deleted file mode 100644
index f21147ef..00000000
--- a/ansible/playbooks/paas/roles/coredns/templates/prerm.j2
+++ /dev/null
@@ -1,13 +0,0 @@
-#!/bin/bash
-
-set -e
-
-# Variables
-SERVICE_NAME="{{ image.name }}"
-
-# Stop and disable the service
-if command -v systemctl >/dev/null 2>&1; then
-  echo "Stopping and disabling $SERVICE_NAME service..."
-  systemctl stop "$SERVICE_NAME" || true
-  systemctl disable "$SERVICE_NAME" || true
-fi
diff --git a/ansible/playbooks/paas/roles/coredns/vars/main.yml b/ansible/playbooks/paas/roles/coredns/vars/main.yml
index 11175960..262e254c 100644
--- a/ansible/playbooks/paas/roles/coredns/vars/main.yml
+++ b/ansible/playbooks/paas/roles/coredns/vars/main.yml
@@ -3,14 +3,14 @@ image:
   build: false
   upstream:
     source: github
-    user: ituoga
-    repo: coredns-nomad
-    type: archive
-    format: tar.gz
-    file: VERSION.FORMAT
+    user: coredns
+    repo: coredns
+    type: release
+    format: tgz
+    file: coredns_VERSION_OS_ARCH.FORMAT
     os: linux
     binary: coredns
   labels: {}
   name: coredns
 
-build_work_dir: /tmp/coredns-nomad
+build_work_dir: /tmp/coredns
diff --git a/ansible/playbooks/paas/roles/coredns/vars/upstream.yml b/ansible/playbooks/paas/roles/coredns/vars/upstream.yml
index f6893ce2..036aa64c 100644
--- a/ansible/playbooks/paas/roles/coredns/vars/upstream.yml
+++ b/ansible/playbooks/paas/roles/coredns/vars/upstream.yml
@@ -1,4 +1,4 @@
 ---
 latest_version: "{{ (lookup('url', 'https://api.github.com/repos/{{ image.upstream.user }}/{{ image.upstream.repo }}/releases/latest', headers={'Accept': 'application/vnd.github+json', 'Authorization': 'Bearer ' + lookup('ansible.builtin.env', 'GITHUB_API_TOKEN') }) | from_json).get('tag_name') | replace('v', '') }}"
 upstream_file_name: "{{ image.upstream.file | replace('REPO', image.upstream.repo) | replace('VERSION', latest_version) | replace('OS', image.upstream.os) | replace('ARCH', upstream_default_arch) | replace('FORMAT', image.upstream.format) }}"
-upstream_file_url: "https://github.com/{{ image.upstream.user }}/{{ image.upstream.repo }}/archive/refs/tags/v{{ upstream_file_name }}"
+upstream_file_url: "https://github.com/{{ image.upstream.user }}/{{ image.upstream.repo }}/releases/download/v{{ latest_version }}/{{ upstream_file_name }}"
diff --git a/ansible/playbooks/paas/roles/node_exporter/templates/default.j2 b/ansible/playbooks/paas/roles/node_exporter/templates/default.j2
index acb52c40..1289c4c7 100644
--- a/ansible/playbooks/paas/roles/node_exporter/templates/default.j2
+++ b/ansible/playbooks/paas/roles/node_exporter/templates/default.j2
@@ -1,5 +1,5 @@
 ARGS="--log.level=info \
---web.listen-address=127.0.0.1:9100 \
+--web.listen-address={{ hostvars[inventory_hostname]['ansible_' + nomad_iface]['ipv4']['address'] }}:9100 \
 --web.telemetry-path=/metrics \
 --collector.diskstats.ignored-devices='^(ram|loop|fd|(h|s|v|xv)d[a-z]|nbd|nvme[0-9]+n[0-9]+p|md|dm-)[0-9]+$' \
 --collector.filesystem.mount-points-exclude='^/(dev(/shm)?|proc|run(/.+)?|sys|var/tmp|(var/lib|home)/(docker|kubelet)/.+)($|/)' \
diff --git a/ansible/playbooks/paas/roles/nomad/defaults/main.yml b/ansible/playbooks/paas/roles/nomad/defaults/main.yml
index b4945a05..86c61ac5 100644
--- a/ansible/playbooks/paas/roles/nomad/defaults/main.yml
+++ b/ansible/playbooks/paas/roles/nomad/defaults/main.yml
@@ -15,7 +15,7 @@ nomad_timezone: "Europe/Paris"
 nomad_group: simplestack
 
 # Configuration
-nomad_dc_name: "dc1"
+nomad_dc_name: dc1
 
 nomad_project: "{{ fact_instance.project }}"
 nomad_region: "{{ fact_instance.region }}"
@@ -43,15 +43,15 @@ nomad_leave_on_interrupt: true
 nomad_client_auto_join: true
 nomad_server_auto_join: true
 
-nomad_s3_storage_enabled: true
+nomad_s3_storage_enabled: false
 
 # Network
 nomad_http_scheme: https
 nomad_http_ip: "127.0.0.1"
 nomad_http_port: 4646
 
-nomad_cluster_bridge: "ens3"
-nomad_iface: "ens3"
+nomad_cluster_bridge: ens3
+nomad_iface: ens3
 
 nomad_bind_address: "0.0.0.0"
 nomad_advertise_address: "{{ hostvars[inventory_hostname]['ansible_' + nomad_iface]['ipv4']['address'] }}"
@@ -146,25 +146,38 @@ nomad_client_host_network_cluster:
   name: cluster
   interface: "{{ nomad_cluster_bridge }}"
 
-nomad_client_meta_list: {"arch": "{{ architecture_map[ansible_facts.architecture] }}", "location": "{{ fact_instance.location }}", "instance": "{{ inventory_hostname }}"}
+nomad_client_meta_list: >-
+  {"arch": "{{ architecture_map[ansible_facts.architecture] }}",
+   "location": "{{ fact_instance.location }}",
+   "instance": "{{ inventory_hostname }}"}
+
+nomad_server_join: >-
+  "{% if nomad_mode == 'single' %}127.0.0.1{% else %}{{ (groups[nomad_deploy_cluster_name] |
+  map('extract', hostvars) |
+  selectattr('nomad_node_role', 'equalto', 'both') |
+  map(attribute='ansible_br0.ipv4.address')) or
+  (groups[nomad_deploy_cluster_name] |
+  map('extract', hostvars) |
+  selectattr('nomad_node_role', 'equalto', 'both') |
+  map(attribute='ansible_br0.ipv4.address')) |
+  unique | list }}{% endif %}"
 
-nomad_server_join: "{% if nomad_mode == 'single' %}127.0.0.1{% else %}{{ (groups[nomad_deploy_cluster_name] | map('extract', hostvars) | selectattr('nomad_node_role', 'equalto', 'both') | map(attribute='ansible_br0.ipv4.address')) or (groups[nomad_deploy_cluster_name] | map('extract', hostvars) | selectattr('nomad_node_role', 'equalto', 'both') | map(attribute='ansible_br0.ipv4.address')) | unique | list }}{% endif %}"
 nomad_server_join_retry_max: 3
-nomad_server_join_retry_interval: "15s"
+nomad_server_join_retry_interval: 15s
 
 nomad_client_server_join_retry_max: 3
-nomad_client_server_join_retry_interval: "15s"
+nomad_client_server_join_retry_interval: 15s
 
-nomad_client_drain_on_shutdown_deadline: "1m"
-nomad_client_drain_on_shutdown_force: "true"
-nomad_client_drain_on_shutdown_ignore_system_jobs: "true"
+nomad_client_drain_on_shutdown_deadline: 1m
+nomad_client_drain_on_shutdown_force: true
+nomad_client_drain_on_shutdown_ignore_system_jobs: true
 
 nomad_client_cpu_total_compute: 0
 nomad_client_memory_total_mb: 0
 nomad_client_disk_total_mb: 0
 nomad_client_disk_free_mb: 0
 
-nomad_client_gc_interval: "1m"
+nomad_client_gc_interval: 1m
 nomad_client_gc_disk_usage_threshold: 80
 nomad_client_gc_inode_usage_threshold: 70
 nomad_client_gc_parallel_destroys: 2
@@ -175,12 +188,15 @@ nomad_client_reserved_disk: 0
 
 # TLS
 nomad_tls_ca_host: localhost
-nomad_tls_ca_host_dir: "~/.simple-stack/tls"
-nomad_tls_ca_pubkey: "simplestack-ca.pem"
-nomad_tls_ca_privatekey: "simplestack-ca-key.pem"
-nomad_tls_ca_provider: "ownca"
-nomad_tls_host_certificate_dir: "/etc/ssl/simplestack"
-nomad_tls_common_name: "nomad"
+nomad_tls_ca_host_dir: ~/.simple-stack/tls
+nomad_tls_ca_pubkey: simplestack-ca.pem
+nomad_tls_ca_privatekey: simplestack-ca-key.pem
+nomad_tls_ca_provider: ownca
+nomad_tls_host_certificate_dir: /etc/ssl/simplestack
+
+nomad_tls_common_name: nomad
+# IP range for 192.168.0.0/24 (all 256 addresses)
+nomad_tls_ip_range: "{{ range(0,256) | map('regex_replace', '^', 'IP:192.168.0.') | list | join(',') }}"
 nomad_tls_check_delay: "+2w"
 
 # TLS Server
@@ -188,36 +204,32 @@ nomad_tls_cert_server: "{{ nomad_dc_name }}-server-nomad.pem"
 nomad_tls_privatekey_server: "{{ nomad_dc_name }}-server-nomad.key"
 
 nomad_tls_common_name_server: "*.{{ nomad_dc_name }}.{{ nomad_tls_common_name }}"
-# nomad_tls_subject_alt_name_server: "DNS:localhost,IP:127.0.0.1,DNS:server.global.{{ certificate_subject_alt_name }},DNS:server.{{ nomad_region }}.{{ certificate_subject_alt_name }},DNS:server.{{ nomad_dc_name }}.{{ certificate_subject_alt_name }},DNS:*.{{ nomad_dc_name }}.{{ certificate_subject_alt_name }},IP:172.26.64.1,IP:172.17.0.1,IP:172.18.0.1"
-# nomad_tls_subject_alt_name_server: "DNS:localhost,IP:127.0.0.1,DNS:server.global.nomad,DNS:server.{{ nomad_region }}.nomad,DNS:server.{{ nomad_dc_name }}.nomad,DNS:*.{{ nomad_dc_name }}.nomad,IP:172.26.64.1,IP:172.17.0.1,IP:172.18.0.1"
-nomad_tls_subject_alt_name_server: "DNS:localhost,IP:127.0.0.1,IP:172.17.0.1,DNS:server.global.nomad,DNS:server.{{ nomad_region }}.nomad,DNS:server.{{ nomad_dc_name }}.nomad,DNS:*.{{ nomad_dc_name }}.nomad"
+nomad_tls_subject_alt_name_server: "DNS:localhost,IP:127.0.0.1,IP:172.17.0.1,{{ nomad_tls_ip_range }},DNS:server.global.nomad,DNS:server.{{ nomad_region }}.nomad,DNS:server.{{ nomad_dc_name }}.nomad,DNS:*.{{ nomad_dc_name }}.nomad"
 
 # TLS client
 nomad_tls_cert_client: "{{ inventory_hostname }}-{{ nomad_dc_name }}-client-nomad.pem"
 nomad_tls_privatekey_client: "{{ inventory_hostname }}-{{ nomad_dc_name }}-client-nomad.key"
 
 nomad_tls_common_name_client: "*.{{ nomad_dc_name }}.{{ nomad_tls_common_name }}"
-# nomad_tls_subject_alt_name_client: "DNS:localhost,IP:127.0.0.1,DNS:client.global.{{ certificate_subject_alt_name }},DNS:client.{{ nomad_region }}.{{ nomad_tls_common_name }},DNS:client.{{ nomad_dc_name }}.{{ nomad_tls_common_name }},DNS:*.{{ nomad_dc_name }}.{{ nomad_tls_common_name }},IP:172.26.64.1,IP:172.17.0.1,IP:172.18.0.1"
-# nomad_tls_subject_alt_name_client: "DNS:localhost,IP:127.0.0.1,DNS:client.global.nomad,DNS:client.{{ nomad_region }}.nomad,DNS:client.{{ nomad_dc_name }}.nomad,DNS:*.{{ nomad_dc_name }}.nomad,IP:172.26.64.1,IP:172.17.0.1,IP:172.18.0.1"
-nomad_tls_subject_alt_name_client: "DNS:localhost,IP:127.0.0.1,IP:172.17.0.1,DNS:client.global.nomad,DNS:client.{{ nomad_region }}.nomad,DNS:client.{{ nomad_dc_name }}.nomad,DNS:*.{{ nomad_dc_name }}.nomad"
+nomad_tls_subject_alt_name_client: "DNS:localhost,IP:127.0.0.1,IP:172.17.0.1,{{ nomad_tls_ip_range }},DNS:client.global.nomad,DNS:client.{{ nomad_region }}.nomad,DNS:client.{{ nomad_dc_name }}.nomad,DNS:*.{{ nomad_dc_name }}.nomad"
 
-nomad_tls_rpc_upgrade_mode: "false"
-nomad_tls_verify_server_hostname: "true"
-nomad_tls_verify_https_client: "false"
+nomad_tls_rpc_upgrade_mode: false
+nomad_tls_verify_server_hostname: true
+nomad_tls_verify_https_client: false
 
 # ACL
 nomad_acl_enabled: true
 nomad_acl_token_ttl: 30s
 nomad_acl_policy_ttl: 30s
-nomad_acl_replication_token: ""
+nomad_acl_replication_token:
 
 # Docker
-nomad_docker_client_dc_name: "dc1"
+nomad_docker_client_dc_name: "{{ nomad_dc_name }}"
 
 nomad_docker_tcp_listen_address: "127.0.0.1"
 nomad_docker_tcp_listen_port: 2376
 
-docker_tls_configuration: false
+nomad_docker_tls_configuration: false
 
 nomad_docker_client_tls_host_certificate_dir: "/etc/ssl/docker"
 nomad_docker_client_tls_cert: "{{ nomad_docker_client_dc_name }}-client-docker.pem"
@@ -241,40 +253,40 @@ nomad_docker_client_allow_caps:
   - sys_ptrace
   - sys_admin
 
-nomad_docker_private_registry_state: false
+nomad_docker_private_registry_state: true
 nomad_docker_private_registry_config: /etc/docker/config.json
 
-nomad_docker_allow_privileged: "{% if nomad_s3_storage_enabled %}true{% else %}false{% endif %}"
-nomad_docker_volume_enable: "true"
-nomad_docker_gc_image: "true"
-nomad_docker_gc_image_delay: "1h"
-nomad_docker_gc_container: "true"
-nomad_docker_gc_dangling_containers_enabled: "true"
-nomad_docker_gc_dangling_containers_dry_run: "false"
-nomad_docker_gc_dangling_containers_period: "5m"
-nomad_docker_gc_dangling_containers_creation_grace: "5m"
+nomad_docker_allow_privileged: true
+nomad_docker_volume_enable: true
+nomad_docker_gc_image: true
+nomad_docker_gc_image_delay: 1h
+nomad_docker_gc_container: true
+nomad_docker_gc_dangling_containers_enabled: true
+nomad_docker_gc_dangling_containers_dry_run: false
+nomad_docker_gc_dangling_containers_period: 5m
+nomad_docker_gc_dangling_containers_creation_grace: 5m
 
 # Telemetry
-nomad_telemetry_disable_hostname: "false"
-nomad_telemetry_collection_interval: "5s"
-nomad_telemetry_use_node_name: "false"
-nomad_telemetry_publish_allocation_metrics: "true"
-nomad_telemetry_publish_node_metrics: "true"
-nomad_telemetry_filter_default: "true"
+nomad_telemetry_disable_hostname: false
+nomad_telemetry_collection_interval: 5s
+nomad_telemetry_use_node_name: false
+nomad_telemetry_publish_allocation_metrics: true
+nomad_telemetry_publish_node_metrics: true
+nomad_telemetry_filter_default: true
 # nomad_telemetry_prefix_filter:
-nomad_telemetry_disable_dispatched_job_summary_metrics: "false"
+nomad_telemetry_disable_dispatched_job_summary_metrics: false
 # nomad_telemetry_statsite_address: ""
 # nomad_telemetry_statsd_address: ""
 # nomad_telemetry_datadog_address: ""
 # nomad_telemetry_datadog_tags:
-nomad_telemetry_prometheus_metrics: "true"
+nomad_telemetry_prometheus_metrics: true
 # nomad_telemetry_circonus_api_token: ""
 nomad_telemetry_circonus_api_app: "nomad"
 nomad_telemetry_circonus_api_url: "https://api.circonus.com/v2"
 nomad_telemetry_circonus_submission_interval: "10s"
 # nomad_telemetry_circonus_submission_url: ""
 # nomad_telemetry_circonus_check_id: ""
-nomad_telemetry_circonus_check_force_metric_activation: "false"
+nomad_telemetry_circonus_check_force_metric_activation: false
 # nomad_telemetry_circonus_check_instance_id: ""
 # nomad_telemetry_circonus_check_search_tag: ""
 # nomad_telemetry_circonus_check_display_name: ""
@@ -299,5 +311,5 @@ nomad_ui_content_security_policy_script_src: "'self'"
 nomad_ui_content_security_policy_style_src: ""
 
 nomad_ui_label_text: "{{ inventory_hostname }}"
-nomad_ui_label_background_color: "blue"
-nomad_ui_label_text_color: "white"
+nomad_ui_label_background_color: blue
+nomad_ui_label_text_color: white
diff --git a/ansible/playbooks/paas/roles/nomad/tasks/04_tls_certs.yml b/ansible/playbooks/paas/roles/nomad/tasks/04_tls_certs.yml
index 54f88757..43fc09ae 100644
--- a/ansible/playbooks/paas/roles/nomad/tasks/04_tls_certs.yml
+++ b/ansible/playbooks/paas/roles/nomad/tasks/04_tls_certs.yml
@@ -45,24 +45,22 @@
       run_once: true
       when: not cert_tls_server_present.stat.exists or (cert_tls_server_present.stat.exists and not tls_check_server.valid_at.delay)
 
-    - name: "Nomad | Copy cert private server key on nodes"
+    - name: "Nomad | Copy certificates on server nodes"
       ansible.builtin.copy:
-        src: "{{ nomad_tls_ca_host_dir }}/{{ nomad_tls_privatekey_server }}"
-        dest: "{{ nomad_tls_host_certificate_dir }}/{{ nomad_tls_privatekey_server }}"
-        owner: "root"
-        group: "{{ nomad_group }}"
-        mode: "0640"
-
-    - name: "Nomad | Copy cert server on nodes"
-      ansible.builtin.copy:
-        src: "{{ nomad_tls_ca_host_dir }}/{{ nomad_tls_cert_server }}"
-        dest: "{{ nomad_tls_host_certificate_dir }}/{{ nomad_tls_cert_server }}"
-        owner: "root"
+        src: "{{ item.src }}"
+        dest: "{{ item.dest }}"
+        owner: root
         group: "{{ nomad_group }}"
         mode: "0640"
+      loop:
+        - src: "{{ nomad_tls_ca_host_dir }}/{{ nomad_tls_privatekey_server }}"
+          dest: "{{ nomad_tls_host_certificate_dir }}/{{ nomad_tls_privatekey_server }}"
+        - src: "{{ nomad_tls_ca_host_dir }}/{{ nomad_tls_cert_server }}"
+          dest: "{{ nomad_tls_host_certificate_dir }}/{{ nomad_tls_cert_server }}"
+      notify: Nomad_restart
 
 - name: Nomad | Copy certificate on client nodes
-  when: nomad_node_role == 'client'
+  when: nomad_node_role in ['client', 'both']
   block:
     - name: "Nomad | Check if TLS cert exists for Client"
       ansible.builtin.stat:
@@ -90,20 +88,18 @@
         certificate_client_privatekey: "{{ nomad_tls_privatekey_client }}"
         certificate_common_name: "{{ nomad_tls_common_name_client }}"
         certificate_subject_alt_name: "{{ nomad_tls_subject_alt_name_client }}"
-      when: nomad_mode == 'cluster'
-
-    - name: "Nomad | Copy cert client key on nodes"
-      ansible.builtin.copy:
-        src: "{{ nomad_tls_ca_host_dir }}/{{ nomad_tls_cert_client }}"
-        dest: "{{ nomad_tls_host_certificate_dir }}/{{ nomad_tls_cert_client }}"
-        owner: "root"
-        group: "{{ nomad_group }}"
-        mode: "0640"
+      # when: nomad_mode == 'cluster'
 
-    - name: "Nomad | Copy cert private client key on nodes"
+    - name: "Nomad | Copy certificates on client nodes"
       ansible.builtin.copy:
-        src: "{{ nomad_tls_ca_host_dir }}/{{ nomad_tls_privatekey_client }}"
-        dest: "{{ nomad_tls_host_certificate_dir }}/{{ nomad_tls_privatekey_client }}"
-        owner: "root"
+        src: "{{ item.src }}"
+        dest: "{{ item.dest }}"
+        owner: root
         group: "{{ nomad_group }}"
         mode: "0640"
+      loop:
+        - src: "{{ nomad_tls_ca_host_dir }}/{{ nomad_tls_cert_client }}"
+          dest: "{{ nomad_tls_host_certificate_dir }}/{{ nomad_tls_cert_client }}"
+        - src: "{{ nomad_tls_ca_host_dir }}/{{ nomad_tls_privatekey_client }}"
+          dest: "{{ nomad_tls_host_certificate_dir }}/{{ nomad_tls_privatekey_client }}"
+      notify: Nomad_restart
diff --git a/ansible/playbooks/paas/roles/nomad/tasks/06_configuration.yml b/ansible/playbooks/paas/roles/nomad/tasks/06_configuration.yml
index fc3057c1..b32cecc8 100644
--- a/ansible/playbooks/paas/roles/nomad/tasks/06_configuration.yml
+++ b/ansible/playbooks/paas/roles/nomad/tasks/06_configuration.yml
@@ -68,10 +68,11 @@
     port: "{{ nomad_http_port }}"
 
 - name: Block
+  when: nomad_node_role in ['server', 'both']
   block:
     - name: "Nomad Install | Read Nomad management token from UI"
       ansible.builtin.set_fact:
-        nomad_management_token: "{{ lookup('simple-stack-ui', type='secret', key=inventory_hostname, subkey='nomad_management_token', missing='error') }}"
+        nomad_management_token: "{{ lookup('simple-stack-ui', type='secret', key=nomad_primary_master_node | default(inventory_hostname), subkey='nomad_management_token', missing='error') }}"
   rescue:
     - name: "Nomad ACL | Generate Bootstrap token"
       ansible.builtin.uri:
@@ -88,7 +89,7 @@
 
     - name: "Nomad Install | Set Nomad management token and insert in UI"
       ansible.builtin.set_fact:
-        nomad_management_token: "{{ lookup('simple-stack-ui', type='secret', key=inventory_hostname, subkey='nomad_management_token', missing='create', userpass=nomad_management_token_result.json.SecretID) }}"
+        nomad_management_token: "{{ lookup('simple-stack-ui', type='secret', key=nomad_primary_master_node | default(inventory_hostname), subkey='nomad_management_token', missing='create', userpass=nomad_management_token_result.json.SecretID) }}"
 
 - name: "Nomad Configuration | Enable MemoryOversubscription"
   ansible.builtin.uri:
@@ -109,18 +110,18 @@
   register: nomad_memoryoversubscription
   ignore_errors: true
 
-- name: "Nomad Configuration | Add S3 storage plugin job templates"
-  ansible.builtin.template:
-    src: "{{ item }}.j2"
-    dest: "{{ nomad_job_files_dir }}/{{ item }}"
-    mode: '0644'
-  loop:
-    - "plugin-s3-controller.hcl"
-    - "plugin-s3-node.hcl"
-  when:
-    - nomad_s3_storage_enabled
-    - nomad_node_role in ['client', 'both']
-  notify: Nomad_s3_jobs
+# - name: "Nomad Configuration | Add S3 storage plugin job templates"
+#   ansible.builtin.template:
+#     src: "{{ item }}.j2"
+#     dest: "{{ nomad_job_files_dir }}/{{ item }}"
+#     mode: '0644'
+#   loop:
+#     - "plugin-s3-controller.hcl"
+#     - "plugin-s3-node.hcl"
+#   when:
+#     - nomad_s3_storage_enabled
+#     - nomad_node_role in ['client', 'both']
+#   notify: Nomad_s3_jobs
 
-- name: "Nomad Configuration | Flush handlers"
-  ansible.builtin.meta: flush_handlers
+# - name: "Nomad Configuration | Flush handlers"
+#   ansible.builtin.meta: flush_handlers
diff --git a/ansible/playbooks/paas/roles/nomad/tasks/10_juicefs.yml b/ansible/playbooks/paas/roles/nomad/tasks/10_juicefs.yml
new file mode 100644
index 00000000..d5414e97
--- /dev/null
+++ b/ansible/playbooks/paas/roles/nomad/tasks/10_juicefs.yml
@@ -0,0 +1,59 @@
+---
+- name: "Nomad Juicefs | Copy configuration"
+  ansible.builtin.template:
+    src: "{{ item }}.j2"
+    dest: "{{ nomad_job_files_dir }}/{{ item }}"
+    mode: '0644'
+  loop:
+    - juicefs-controller.hcl
+    - juicefs-node.hcl
+    - juicefs-volume.hcl
+
+- name: "Nomad Juicefs | Copy volume configuration"
+  ansible.builtin.template:
+    src: juicefs-volume.hcl.j2
+    dest: "{{ nomad_job_files_dir }}/juicefs-volume-{{ item }}.hcl"
+    mode: '0644'
+  loop:
+    - volume
+    - test
+    - llm
+
+- name: Nomad Juicefs | Run jobs
+  ansible.builtin.command: "/usr/bin/nomad job run {{ nomad_job_files_dir }}/{{ item }}"
+  args:
+    chdir: "{{ nomad_job_files_dir }}"
+  environment:
+    NOMAD_ADDR: "https://{{ nomad_http_ip }}:4646"
+    NOMAD_TOKEN: "{{ lookup('simple-stack-ui', type='secret', key=inventory_hostname, subkey='nomad_management_token', missing='error') }}"
+    NOMAD_CLIENT_CERT: "{{ nomad_tls_host_certificate_dir }}/{{ nomad_tls_cert_server }}"
+    NOMAD_CLIENT_KEY: "{{ nomad_tls_host_certificate_dir }}/{{ nomad_tls_privatekey_server }}"
+    NOMAD_CACERT: "{{ nomad_tls_host_certificate_dir }}/{{ nomad_tls_ca_pubkey }}"
+  register: nomad_job_start
+  loop:
+    - juicefs-controller.hcl
+    - juicefs-node.hcl
+  failed_when: nomad_job_start.rc >= 2
+  changed_when:
+    - '"error" in nomad_job_start.stdout'
+    - nomad_job_start.rc >= 2
+
+- name: Nomad Juicefs | Create volume
+  ansible.builtin.command: "/usr/bin/nomad volume create {{ nomad_job_files_dir }}/juicefs-volume-{{ item }}.hcl"
+  args:
+    chdir: "{{ nomad_job_files_dir }}"
+  environment:
+    NOMAD_ADDR: "https://{{ nomad_http_ip }}:4646"
+    NOMAD_TOKEN: "{{ lookup('simple-stack-ui', type='secret', key=inventory_hostname, subkey='nomad_management_token', missing='error') }}"
+    NOMAD_CLIENT_CERT: "{{ nomad_tls_host_certificate_dir }}/{{ nomad_tls_cert_server }}"
+    NOMAD_CLIENT_KEY: "{{ nomad_tls_host_certificate_dir }}/{{ nomad_tls_privatekey_server }}"
+    NOMAD_CACERT: "{{ nomad_tls_host_certificate_dir }}/{{ nomad_tls_ca_pubkey }}"
+  register: nomad_job_start
+  loop:
+    - volume
+    - test
+    - llm
+  failed_when: nomad_job_start.rc >= 2
+  changed_when:
+    - '"error" in nomad_job_start.stdout'
+    - nomad_job_start.rc >= 2
diff --git a/ansible/playbooks/paas/roles/nomad/tasks/main.yml b/ansible/playbooks/paas/roles/nomad/tasks/main.yml
index 75a72cd8..51957285 100644
--- a/ansible/playbooks/paas/roles/nomad/tasks/main.yml
+++ b/ansible/playbooks/paas/roles/nomad/tasks/main.yml
@@ -7,8 +7,8 @@
     name: "{{ nomad_timezone }}"
     hwclock: local
 
-- name: "Nomad | Install CNI"
-  ansible.builtin.include_tasks: "02_network.yml"
+# - name: "Nomad | Install CNI"
+#   ansible.builtin.include_tasks: "02_network.yml"
 
 - name: "Nomad | Commons tasks"
   ansible.builtin.include_tasks: "03_commons_tasks.yml"
@@ -27,4 +27,3 @@
 
 - name: "Nomad | Change SystemD configuration"
   ansible.builtin.include_tasks: "08_systemd_tuning.yml"
-
diff --git a/ansible/playbooks/paas/roles/nomad/templates/client.hcl.j2 b/ansible/playbooks/paas/roles/nomad/templates/client.hcl.j2
index 184a5b9d..a7db1122 100644
--- a/ansible/playbooks/paas/roles/nomad/templates/client.hcl.j2
+++ b/ansible/playbooks/paas/roles/nomad/templates/client.hcl.j2
@@ -1,12 +1,12 @@
 client {
-    enabled = {{ nomad_client_enabled | bool | lower }}
+    enabled = {{ nomad_client_enabled | lower }}
 
     state_dir = "{{ nomad_state_dir_client }}"
 
     node_class = "{{ nomad_client_node_class }}"
     node_pool = "{{ nomad_client_node_pool }}"
 
-    no_host_uuid = {{ nomad_client_no_host_uuid | bool | lower }}
+    no_host_uuid = {{ nomad_client_no_host_uuid | lower }}
 
     servers = [
         {%- set comma = joiner(",") -%}
@@ -51,8 +51,8 @@ client {
 {% if nomad_mode == 'cluster' %}
     drain_on_shutdown {
       deadline           = "{{ nomad_client_drain_on_shutdown_deadline }}"
-      force              = {{ nomad_client_drain_on_shutdown_force }}
-      ignore_system_jobs = {{ nomad_client_drain_on_shutdown_ignore_system_jobs }}
+      force              = {{ nomad_client_drain_on_shutdown_force | lower }}
+      ignore_system_jobs = {{ nomad_client_drain_on_shutdown_ignore_system_jobs | lower }}
     }
 {% endif %}
 
diff --git a/ansible/playbooks/paas/roles/nomad/templates/docker.hcl.j2 b/ansible/playbooks/paas/roles/nomad/templates/docker.hcl.j2
index 14043718..a82ce90c 100644
--- a/ansible/playbooks/paas/roles/nomad/templates/docker.hcl.j2
+++ b/ansible/playbooks/paas/roles/nomad/templates/docker.hcl.j2
@@ -1,6 +1,6 @@
 plugin "docker" {
     config {
-{% if docker_tls_configuration == true %}
+{% if nomad_docker_tls_configuration %}
     	endpoint = "tcp://{{ nomad_docker_tcp_listen_address }}:{{ nomad_docker_tcp_listen_port }}"
 
     	tls {
@@ -12,27 +12,27 @@ plugin "docker" {
     	endpoint = "unix:///var/run/docker.sock"
 {% endif %}
 
-{% if nomad_docker_private_registry_state == true %}
+{% if nomad_docker_private_registry_state %}
         auth {
             config = "{{ nomad_docker_private_registry_config }}"
         }
 {% endif %}
 
-        allow_privileged = {{ nomad_docker_allow_privileged }}
+        allow_privileged = {{ nomad_docker_allow_privileged | lower }}
 
         volumes {
-            enabled = {{ nomad_docker_volume_enable }}
+            enabled = {{ nomad_docker_volume_enable | lower }}
         }
         allow_caps = [{% for item in nomad_docker_client_allow_caps %}"{{ item }}"{% if not loop.last %}, {% endif %}{% endfor %}]
 
         gc {
-            image       = {{ nomad_docker_gc_image }}
-            image_delay = "{{ nomad_docker_gc_image_delay }}"
-            container   = {{ nomad_docker_gc_container }}
+            image       = {{ nomad_docker_gc_image | lower }}
+            image_delay = "{{ nomad_docker_gc_image_delay | lower }}"
+            container   = {{ nomad_docker_gc_container | lower }}
 
             dangling_containers {
-                enabled        = {{ nomad_docker_gc_dangling_containers_enabled }}
-                dry_run        = {{ nomad_docker_gc_dangling_containers_dry_run }}
+                enabled        = {{ nomad_docker_gc_dangling_containers_enabled | lower }}
+                dry_run        = {{ nomad_docker_gc_dangling_containers_dry_run | lower }}
                 period         = "{{ nomad_docker_gc_dangling_containers_period }}"
                 creation_grace = "{{ nomad_docker_gc_dangling_containers_creation_grace }}"
             }
diff --git a/ansible/playbooks/paas/roles/nomad/templates/juicefs-controller.hcl.j2 b/ansible/playbooks/paas/roles/nomad/templates/juicefs-controller.hcl.j2
new file mode 100644
index 00000000..c42da612
--- /dev/null
+++ b/ansible/playbooks/paas/roles/nomad/templates/juicefs-controller.hcl.j2
@@ -0,0 +1,43 @@
+job "jfs-controller" {
+  datacenters = ["dc1"]
+  type = "system"
+
+  group "controller" {
+    
+    constraint {
+        attribute    = "${meta.instance}"
+        set_contains = "{{ nomad_constraints_juicefs_controller_instance }}"
+    }
+
+    task "plugin" {
+      driver = "docker"
+
+      config {
+        image = "juicedata/juicefs-csi-driver:v0.30.0"
+
+        args = [
+          "--endpoint=unix://csi/csi.sock",
+          "--logtostderr",
+          "--nodeid=test",
+          "--v=5",
+          "--by-process=true"
+        ]
+
+        privileged = true
+      }
+
+      csi_plugin {
+        id        = "juicefs0"
+        type      = "controller"
+        mount_dir = "/csi"
+      }
+      resources {
+        cpu    = 100
+        memory = 512
+      }
+      env {
+        POD_NAME = "csi-controller"
+      }
+    }
+  }
+}
\ No newline at end of file
diff --git a/ansible/playbooks/paas/roles/nomad/templates/juicefs-node.hcl.j2 b/ansible/playbooks/paas/roles/nomad/templates/juicefs-node.hcl.j2
new file mode 100644
index 00000000..f503ef6d
--- /dev/null
+++ b/ansible/playbooks/paas/roles/nomad/templates/juicefs-node.hcl.j2
@@ -0,0 +1,48 @@
+job "jfs-node" {
+  datacenters = ["dc1"]
+  type = "system"
+
+  group "nodes" {
+
+    constraint {
+        attribute = "${meta.location}"
+        operator  = "set_contains_any"
+        value     = "{{ nomad_constraints_juicefs_controller_nodes | join(',') }}"
+    }
+
+    task "juicefs-plugin" {
+      driver = "docker"
+
+      config {
+        image = "juicedata/juicefs-csi-driver:v0.30.0"
+
+        args = [
+          "--endpoint=unix://csi/csi.sock",
+          "--logtostderr",
+          "--v=5",
+          "--nodeid=test",
+          "--by-process=true",
+          #"--cache-dir", "/var/jfs-cache",
+          #"--cache-size", "50G",
+          #"--writeback",
+          #"--prefetch", "2"
+        ]
+
+        privileged = true
+      }
+
+      csi_plugin {
+        id        = "juicefs0"
+        type      = "node"
+        mount_dir = "/csi"
+      }
+      resources {
+        cpu    = 1000
+        memory = 1024
+      }
+      env {
+        POD_NAME = "csi-node"
+      }
+    }
+  }
+}
\ No newline at end of file
diff --git a/ansible/playbooks/paas/roles/nomad/templates/juicefs-volume.hcl.j2 b/ansible/playbooks/paas/roles/nomad/templates/juicefs-volume.hcl.j2
new file mode 100644
index 00000000..f45cf80f
--- /dev/null
+++ b/ansible/playbooks/paas/roles/nomad/templates/juicefs-volume.hcl.j2
@@ -0,0 +1,18 @@
+type = "csi"
+id = "juicefs-{{ item }}"
+name = "juicefs-{{ item }}"
+
+capability {
+  access_mode = "multi-node-multi-writer"
+  attachment_mode = "file-system"
+}
+plugin_id = "juicefs0"
+
+secrets {
+  name="juicefs-volume"
+  metaurl="redis://{{ nomad_juicefs_secrets.valkey.address }}:{{ nomad_juicefs_secrets.valkey.port }}/0"
+  bucket="http://{{ nomad_juicefs_secrets.minio.address }}:{{ nomad_juicefs_secrets.minio.port }}/minio/{{ item }}"
+  storage="minio"
+  access-key="{{ lookup('simple-stack-ui', type='secret', key=nomad_juicefs_secrets.minio.domain, subkey='user', missing='error') }}"
+  secret-key="{{ lookup('simple-stack-ui', type='secret', key=nomad_juicefs_secrets.minio.domain, subkey='passwd', missing='error') }}"
+}
\ No newline at end of file
diff --git a/ansible/playbooks/paas/roles/nomad/templates/nomad.hcl.j2 b/ansible/playbooks/paas/roles/nomad/templates/nomad.hcl.j2
index 7d027483..c67aa932 100644
--- a/ansible/playbooks/paas/roles/nomad/templates/nomad.hcl.j2
+++ b/ansible/playbooks/paas/roles/nomad/templates/nomad.hcl.j2
@@ -2,8 +2,8 @@ name = "{{ nomad_node_name }}"
 region = "{{ nomad_region }}"
 datacenter = "{{ nomad_dc_name }}"
 
-disable_anonymous_signature = {{ nomad_disable_anonymous_signature | bool | lower }}
-disable_update_check = {{ nomad_disable_update_check | bool | lower }}
+disable_anonymous_signature = {{ nomad_disable_anonymous_signature | lower }}
+disable_update_check = {{ nomad_disable_update_check | lower }}
 
 data_dir = "{{ nomad_data_dir }}"
 
@@ -20,15 +20,15 @@ ports {
     serf = {{ nomad_ports.serf }}
 }
 
-enable_debug = {{ nomad_debug | bool | lower }}
+enable_debug = {{ nomad_debug | lower }}
 log_file = "{{ nomad_log_file }}"
 log_level = "{{ nomad_log_level }}"
 log_rotate_bytes = {{ nomad_log_rotate_bytes }}
 log_rotate_duration = "{{ nomad_log_rotate_duration }}"
 log_rotate_max_files = {{ nomad_log_rotate_max_files }}
 
-leave_on_terminate = {{ nomad_leave_on_terminate | bool | lower }}
-leave_on_interrupt = {{ nomad_leave_on_interrupt | bool | lower }}
+leave_on_terminate = {{ nomad_leave_on_terminate | lower }}
+leave_on_interrupt = {{ nomad_leave_on_interrupt | lower }}
 
 tls {
     http = true
@@ -36,39 +36,39 @@ tls {
     ca_file = "{{ nomad_tls_host_certificate_dir }}/{{ nomad_tls_ca_pubkey }}"
     cert_file = "{{ nomad_tls_host_certificate_dir }}/{{ (nomad_node_role == 'client') | ternary(nomad_tls_cert_client, nomad_tls_cert_server) }}"
     key_file = "{{ nomad_tls_host_certificate_dir }}/{{ (nomad_node_role == 'client') | ternary(nomad_tls_privatekey_client, nomad_tls_privatekey_server) }}"
-    rpc_upgrade_mode = {{ nomad_tls_rpc_upgrade_mode }}
-    verify_server_hostname = "{{ nomad_tls_verify_server_hostname }}"
-    verify_https_client = "{{ nomad_tls_verify_https_client }}"
+    rpc_upgrade_mode = {{ nomad_tls_rpc_upgrade_mode | lower }}
+    verify_server_hostname = {{ nomad_tls_verify_server_hostname | lower }}
+    verify_https_client = {{ nomad_tls_verify_https_client | lower }}
 }
 
 acl {
-    enabled = {{ nomad_acl_enabled | bool | lower }}
+    enabled = {{ nomad_acl_enabled | lower }}
     token_ttl = "{{ nomad_acl_token_ttl }}"
     policy_ttl = "{{ nomad_acl_policy_ttl }}"
     replication_token = "{{ nomad_acl_replication_token }}"
 }
 
 telemetry {
-    disable_hostname = {{ nomad_telemetry_disable_hostname }}
+    disable_hostname = {{ nomad_telemetry_disable_hostname | lower }}
     collection_interval = "{{ nomad_telemetry_collection_interval }}"
-    use_node_name = {{ nomad_telemetry_use_node_name }}
-    publish_allocation_metrics = {{ nomad_telemetry_publish_allocation_metrics }}
-    publish_node_metrics = {{ nomad_telemetry_publish_node_metrics }}
-    filter_default = {{ nomad_telemetry_filter_default }}
+    use_node_name = {{ nomad_telemetry_use_node_name | lower }}
+    publish_allocation_metrics = {{ nomad_telemetry_publish_allocation_metrics | lower }}
+    publish_node_metrics = {{ nomad_telemetry_publish_node_metrics | lower }}
+    filter_default = {{ nomad_telemetry_filter_default | lower }}
     prefix_filter = []
-    disable_dispatched_job_summary_metrics = {{ nomad_telemetry_disable_dispatched_job_summary_metrics }}
+    disable_dispatched_job_summary_metrics = {{ nomad_telemetry_disable_dispatched_job_summary_metrics | lower }}
     statsite_address = ""
     statsd_address = ""
     datadog_address = ""
     datadog_tags = []
-    prometheus_metrics = {{ nomad_telemetry_prometheus_metrics }}
+    prometheus_metrics = {{ nomad_telemetry_prometheus_metrics | lower }}
     circonus_api_token = ""
     circonus_api_app = "{{ nomad_telemetry_circonus_api_app }}"
     circonus_api_url = "{{ nomad_telemetry_circonus_api_url }}"
     circonus_submission_interval = "{{ nomad_telemetry_circonus_submission_interval }}"
     circonus_submission_url = ""
     circonus_check_id = ""
-    circonus_check_force_metric_activation = {{ nomad_telemetry_circonus_check_force_metric_activation }}
+    circonus_check_force_metric_activation = {{ nomad_telemetry_circonus_check_force_metric_activation | lower }}
     circonus_check_instance_id = ""
     circonus_check_search_tag = ""
     circonus_check_display_name = ""
@@ -78,7 +78,7 @@ telemetry {
 }
 
 autopilot {
-    cleanup_dead_servers      = {{ nomad_autopilot_cleanup_dead_servers | bool | lower }}
+    cleanup_dead_servers      = {{ nomad_autopilot_cleanup_dead_servers | lower }}
     last_contact_threshold    = "{{ nomad_autopilot_last_contact_threshold }}"
     max_trailing_logs         = {{ nomad_autopilot_max_trailing_logs }}
     server_stabilization_time = "{{ nomad_autopilot_server_stabilization_time }}"
@@ -90,7 +90,7 @@ limits {
 }
 
 ui {
-  enabled =  {{ nomad_ui_enabled | bool | lower }}
+  enabled =  {{ nomad_ui_enabled | lower }}
 
   content_security_policy {
     connect_src     = ["{{ nomad_ui_content_security_policy_connect_src }}"]
diff --git a/ansible/playbooks/paas/roles/nomad/templates/server.hcl.j2 b/ansible/playbooks/paas/roles/nomad/templates/server.hcl.j2
index e9edfca8..7a8e8e17 100644
--- a/ansible/playbooks/paas/roles/nomad/templates/server.hcl.j2
+++ b/ansible/playbooks/paas/roles/nomad/templates/server.hcl.j2
@@ -1,11 +1,11 @@
 server {
-    enabled = {{ nomad_server_enabled | bool | lower }}
+    enabled = {{ nomad_server_enabled | lower }}
 
     bootstrap_expect = {{ nomad_servers | length }}
 
     data_dir = "{{ nomad_data_dir_server }}"
 
-    {% if nomad_server_retry_join | bool -%}
+    {% if nomad_server_retry_join -%}
         retry_join = [
             {%- set comma = joiner(",") -%}
             {% for server in nomad_servers_advertise_address -%}
@@ -21,7 +21,7 @@ server {
             {%- endfor -%} ]
     {%- endif %}
 
-    rejoin_after_leave = {{ nomad_server_rejoin_after_leave | bool | lower }}
+    rejoin_after_leave = {{ nomad_server_rejoin_after_leave | lower }}
 
     enabled_schedulers = [
             {%- set comma = joiner(",") -%}
diff --git a/ansible/playbooks/paas/roles/nvidia_gpu_exporter/README.md b/ansible/playbooks/paas/roles/nvidia_gpu_exporter/README.md
new file mode 100644
index 00000000..73adf688
--- /dev/null
+++ b/ansible/playbooks/paas/roles/nvidia_gpu_exporter/README.md
@@ -0,0 +1 @@
+# Role: `nvidia_gpu_exporter`
diff --git a/ansible/playbooks/paas/roles/nvidia_gpu_exporter/defaults/main.yml b/ansible/playbooks/paas/roles/nvidia_gpu_exporter/defaults/main.yml
new file mode 100644
index 00000000..bc455e64
--- /dev/null
+++ b/ansible/playbooks/paas/roles/nvidia_gpu_exporter/defaults/main.yml
@@ -0,0 +1,2 @@
+---
+nvidia_gpu_exporter_enable: false
diff --git a/ansible/playbooks/paas/roles/nvidia_gpu_exporter/handlers/main.yml b/ansible/playbooks/paas/roles/nvidia_gpu_exporter/handlers/main.yml
new file mode 100644
index 00000000..a017bade
--- /dev/null
+++ b/ansible/playbooks/paas/roles/nvidia_gpu_exporter/handlers/main.yml
@@ -0,0 +1,7 @@
+---
+- name: Restart nvidia_gpu_exporter
+  listen: Restart nvidia_gpu_exporter
+  ansible.builtin.service:
+    name: nvidia_gpu_exporter
+    state: restarted
+    enabled: true
diff --git a/ansible/playbooks/paas/roles/nvidia_gpu_exporter/tasks/build.yml b/ansible/playbooks/paas/roles/nvidia_gpu_exporter/tasks/build.yml
new file mode 100644
index 00000000..2dacd7a4
--- /dev/null
+++ b/ansible/playbooks/paas/roles/nvidia_gpu_exporter/tasks/build.yml
@@ -0,0 +1,57 @@
+---
+- name: Nvidia_gpu_exporter | Create temporary build directory
+  ansible.builtin.file:
+    path: "{{ item }}"
+    recurse: true
+    state: directory
+    mode: '0755'
+  loop:
+    - "{{ build_work_dir }}/download"
+    - "{{ build_work_dir }}/{{ upstream_default_arch }}"
+
+- name: Nvidia_gpu_exporter | Download Github release
+  ansible.builtin.get_url:
+    url: "{{ upstream_file_url }}"
+    dest: "{{ build_work_dir }}/download/"
+    mode: '0644'
+  register: download_result
+
+- name: Nvidia_gpu_exporter | Unarchive GitHub release
+  ansible.builtin.unarchive:
+    src: "{{ build_work_dir }}/download/{{ upstream_file_name }}"
+    dest: "{{ build_work_dir }}/download"
+    remote_src: true
+  when: download_result.changed
+
+- name: Nvidia_gpu_exporter | Find binary
+  ansible.builtin.include_role:
+    name: upstream
+    tasks_from: find-binary
+  loop:
+    - "{{ image.upstream.binary }}"
+
+- name: Nvidia_gpu_exporter | Copy binary
+  ansible.builtin.copy:
+    src: "{{ build_work_dir }}/{{ upstream_default_arch }}/{{ image.upstream.binary }}"
+    dest: /usr/local/bin/nvidia_gpu_exporter
+    owner: root
+    group: root
+    mode: '0755'
+    remote_src: true
+
+- name: Nvidia_gpu_exporter | Clean up
+  ansible.builtin.file:
+    path: "{{ build_work_dir }}"
+    state: absent
+
+- name: Nvidia_gpu_exporter | Backup software version
+  ansible.builtin.copy:
+    content: |
+      #!/bin/bash
+      cat << EOF
+      "{{ latest_version }}"
+      EOF
+    dest: "/etc/ansible/facts.d/{{ image.name }}.fact"
+    owner: root
+    group: root
+    mode: '0755'
diff --git a/ansible/playbooks/paas/roles/nvidia_gpu_exporter/tasks/main.yml b/ansible/playbooks/paas/roles/nvidia_gpu_exporter/tasks/main.yml
new file mode 100644
index 00000000..8500b652
--- /dev/null
+++ b/ansible/playbooks/paas/roles/nvidia_gpu_exporter/tasks/main.yml
@@ -0,0 +1,23 @@
+---
+- name: End the play for hosts that don't have nvidia gpu
+  ansible.builtin.meta: end_host
+  when: not nvidia_gpu_exporter_enable
+
+- name: Nvidia_gpu_exporter | Include upstream variables
+  ansible.builtin.include_vars: upstream.yml
+
+- name: Nvidia_gpu_exporter | Get binary
+  ansible.builtin.include_tasks: build.yml
+  when: ansible_local[image.name] is not defined or ansible_local[image.name] != latest_version
+
+- name: Nvidia_gpu_exporter | Copy templates
+  ansible.builtin.template:
+    src: "{{ item.src }}"
+    dest: "{{ item.dest }}"
+    mode: '0644'
+    owner: prometheus
+    group: prometheus
+  loop:
+    - src: service.j2
+      dest: /etc/systemd/system/nvidia_gpu_exporter.service
+  notify: Restart nvidia_gpu_exporter
diff --git a/ansible/playbooks/paas/roles/nvidia_gpu_exporter/templates/service.j2 b/ansible/playbooks/paas/roles/nvidia_gpu_exporter/templates/service.j2
new file mode 100644
index 00000000..1b213d68
--- /dev/null
+++ b/ansible/playbooks/paas/roles/nvidia_gpu_exporter/templates/service.j2
@@ -0,0 +1,16 @@
+[Unit]
+Description=Nvidia GPU Exporter
+Documentation=https://github.com/utkuozdemir/nvidia_gpu_exporter
+After=network-online.target
+
+[Service]
+Type=simple
+User=prometheus
+Group=prometheus
+ExecStart=/usr/local/bin/nvidia_gpu_exporter
+SyslogIdentifier=nvidia_gpu_exporter
+Restart=always
+RestartSec=10
+
+[Install]
+WantedBy=multi-user.target
\ No newline at end of file
diff --git a/ansible/playbooks/paas/roles/nvidia_gpu_exporter/vars/main.yml b/ansible/playbooks/paas/roles/nvidia_gpu_exporter/vars/main.yml
new file mode 100644
index 00000000..8b1563b4
--- /dev/null
+++ b/ansible/playbooks/paas/roles/nvidia_gpu_exporter/vars/main.yml
@@ -0,0 +1,23 @@
+---
+image:
+  build: false
+  upstream:
+    source: github
+    user: utkuozdemir
+    repo: nvidia_gpu_exporter
+    type: release
+    format: tar.gz
+    file: nvidia_gpu_exporter_VERSION_OS_ARCH.FORMAT
+    os: linux
+    binary: nvidia_gpu_exporter
+  labels: {}
+  name: nginx_exporter
+
+build_work_dir: /tmp/nvidia_gpu_exporter
+
+architecture_map:
+  amd64: amd64
+  x86_64: x86_64
+  armv7l: arm
+  aarch64: arm64
+  arm64: arm64
diff --git a/ansible/playbooks/paas/roles/nvidia_gpu_exporter/vars/upstream.yml b/ansible/playbooks/paas/roles/nvidia_gpu_exporter/vars/upstream.yml
new file mode 100644
index 00000000..036aa64c
--- /dev/null
+++ b/ansible/playbooks/paas/roles/nvidia_gpu_exporter/vars/upstream.yml
@@ -0,0 +1,4 @@
+---
+latest_version: "{{ (lookup('url', 'https://api.github.com/repos/{{ image.upstream.user }}/{{ image.upstream.repo }}/releases/latest', headers={'Accept': 'application/vnd.github+json', 'Authorization': 'Bearer ' + lookup('ansible.builtin.env', 'GITHUB_API_TOKEN') }) | from_json).get('tag_name') | replace('v', '') }}"
+upstream_file_name: "{{ image.upstream.file | replace('REPO', image.upstream.repo) | replace('VERSION', latest_version) | replace('OS', image.upstream.os) | replace('ARCH', upstream_default_arch) | replace('FORMAT', image.upstream.format) }}"
+upstream_file_url: "https://github.com/{{ image.upstream.user }}/{{ image.upstream.repo }}/releases/download/v{{ latest_version }}/{{ upstream_file_name }}"
diff --git a/ansible/playbooks/paas/roles/prometheus/tasks/main.yml b/ansible/playbooks/paas/roles/prometheus/tasks/main.yml
index 1911b2de..55e64042 100644
--- a/ansible/playbooks/paas/roles/prometheus/tasks/main.yml
+++ b/ansible/playbooks/paas/roles/prometheus/tasks/main.yml
@@ -33,7 +33,7 @@
   ansible.builtin.template:
     src: "{{ item.src }}"
     dest: "{{ item.dest }}"
-    mode: 0644
+    mode: '0644'
     owner: prometheus
     group: prometheus
   loop:
diff --git a/ansible/playbooks/paas/roles/prometheus/templates/config.j2 b/ansible/playbooks/paas/roles/prometheus/templates/config.j2
index 360d93c9..a4530264 100644
--- a/ansible/playbooks/paas/roles/prometheus/templates/config.j2
+++ b/ansible/playbooks/paas/roles/prometheus/templates/config.j2
@@ -15,9 +15,11 @@ remote_write:
     headers:
       X-Scope-OrgID: demo
     send_exemplars: true
+{% if prometheus_remote_write.basic_auth %}
     basic_auth:
       username: "{{ prometheus_remote_write.login }}"
       password: "{{ prometheus_remote_write.password }}"
+{% endif %}
     queue_config:
       capacity: 25000            # Capacité totale de la file d'attente
       max_shards: 10             # Nombre de shards parallèles (trop haut = surcharge CPU)
@@ -64,14 +66,10 @@ scrape_configs:
         regex: "^(__tmp_keep_me)$"
 
     static_configs:
-      - targets: ['127.0.0.1:9100']
+{% for item in groups['infrastructure'] | default([]) if item.split('.')[4] == project  %}
+      - targets: ['{{ hostvars[item]['ansible_' + hostvars[item].nomad_iface]['ipv4']['address'] }}:9100']
         labels:
-          instance: "{{ inventory_hostname }}"
-          project: "{{ prometheus_project }}"
-{% for item in prometheus_nodes_exporter | default([]) %}
-      - targets: ['{{ item.target }}:9100']
-        labels:
-          instance: "{{ item.instance }}"
+          instance: "{{ item }}"
           project: "{{ prometheus_project }}"
 {% endfor %}
     scrape_interval: 60s
@@ -169,11 +167,39 @@ scrape_configs:
           project: "{{ prometheus_project }}"
 
   - job_name: "systemd_exporter"
+    metric_relabel_configs:
+      - action: drop
+        regex: "^(go_|prometheus_|promhttp_).*"
+        source_labels: [__name__]
+    bearer_token: "{{ lookup('simple-stack-ui', type='secret', key=nomad_primary_master_node | default(inventory_hostname), subkey='nomad_management_token', missing='error') }}"
+    params:
+      format: ['prometheus']
+    metrics_path: /metrics
+    tls_config:
+      insecure_skip_verify: true
     static_configs:
-      - targets: ['127.0.0.1:9558']
+{% for item in groups['infrastructure'] | default([]) if item.split('.')[4] == project %}
+      - targets: ['{{ hostvars[item]['ansible_' + hostvars[item].nomad_iface]['ipv4']['address'] }}:9558']
         labels:
-          instance: "{{ inventory_hostname }}"
+          instance: "{{ item }}"
+          project: "{{ prometheus_project }}"
+{% endfor %}
+
+  - job_name: "nvidia_gpu_exporter"
+    metric_relabel_configs:
+      - action: drop
+        regex: "^(go_|prometheus_|promhttp_).*"
+        source_labels: [__name__]
+    params:
+      format: ['prometheus']
+    metrics_path: /metrics
+    static_configs:
+{% for item in groups['infrastructure'] | default([]) if item.split('.')[4] == project %}
+      - targets: ['{{ hostvars[item]['ansible_' + hostvars[item].nomad_iface]['ipv4']['address'] }}:9835']
+        labels:
+          instance: "{{ item }}"
           project: "{{ prometheus_project }}"
+{% endfor %}
 
   - job_name: "nomad_exporter"
     metric_relabel_configs:
@@ -181,38 +207,51 @@ scrape_configs:
         regex: "^(go_|prometheus_|promhttp_).*"
         source_labels: [__name__]
     scheme: https
-    bearer_token: "{{ lookup('simple-stack-ui', type='secret', key=inventory_hostname, subkey='nomad_management_token', missing='error') }}"
+    bearer_token: "{{ lookup('simple-stack-ui', type='secret', key=nomad_primary_master_node | default(inventory_hostname), subkey='nomad_management_token', missing='error') }}"
     params:
       format: ['prometheus']
     metrics_path: /v1/metrics
     tls_config:
       insecure_skip_verify: true
     static_configs:
-      - targets: ['127.0.0.1:4646']
+{% for item in groups['infrastructure'] | default([]) if item.split('.')[4] == project %}
+      - targets: ['{{ hostvars[item]['ansible_' + hostvars[item].nomad_iface]['ipv4']['address'] }}:4646']
         labels:
-          instance: "{{ inventory_hostname }}"
+          instance: "{{ item }}"
           project: "{{ prometheus_project }}"
-
+{% endfor %}
 
   - job_name: 'mimir_exporter'
     nomad_sd_configs:
-      - server: "https://127.0.0.1:4646"
+      - server: "https://{{ nomad_primary_master_address | default(inventory_hostname) }}:4646"
         region: "{{ fact_instance.region }}"
         tls_config:
           insecure_skip_verify: true
         authorization:
-          credentials: "{{ lookup('simple-stack-ui', type='secret', key=inventory_hostname, subkey='nomad_management_token', missing='error') }}"
+          credentials: "{{ lookup('simple-stack-ui', type='secret', key=nomad_primary_master_node | default(inventory_hostname), subkey='nomad_management_token', missing='error') }}"
     relabel_configs:
       - source_labels: ['__meta_nomad_service']
         regex: 'mimir-exporter'
         action: keep
 
   - job_name: 'promtail'
+    metric_relabel_configs:
+      - action: drop
+        regex: "^(go_|prometheus_|promhttp_).*"
+        source_labels: [__name__]
+    bearer_token: "{{ lookup('simple-stack-ui', type='secret', key=nomad_primary_master_node | default(inventory_hostname), subkey='nomad_management_token', missing='error') }}"
+    params:
+      format: ['prometheus']
+    metrics_path: /metrics
+    tls_config:
+      insecure_skip_verify: true
     static_configs:
-      - targets: ['127.0.0.1:9080']
+{% for item in groups['infrastructure'] | default([]) if item.split('.')[4] == project %}
+      - targets: ['{{ hostvars[item]['ansible_' + hostvars[item].nomad_iface]['ipv4']['address'] }}:9080']
         labels:
-          instance: "{{ inventory_hostname }}"
+          instance: "{{ item }}"
           project: "{{ prometheus_project }}"
+{% endfor %}
 
   - job_name: 'blackbox'
     static_configs:
@@ -223,12 +262,12 @@ scrape_configs:
 
   - job_name: 'traefik'
     nomad_sd_configs:
-      - server: "https://127.0.0.1:4646"
+      - server: "https://{{ nomad_primary_master_address | default(inventory_hostname) }}:4646"
         region: "{{ fact_instance.region }}"
         tls_config:
           insecure_skip_verify: true
         authorization:
-          credentials: "{{ lookup('simple-stack-ui', type='secret', key=inventory_hostname, subkey='nomad_management_token', missing='error') }}"
+          credentials: "{{ lookup('simple-stack-ui', type='secret', key=nomad_primary_master_node | default(inventory_hostname), subkey='nomad_management_token', missing='error') }}"
     relabel_configs:
       - target_label: instance
         replacement: "{{ inventory_hostname }}"
@@ -248,16 +287,22 @@ scrape_configs:
       - source_labels: [__tmp_fqdn]
         target_label: fqdn
 
+      # Traefik exposes its metrics on a different port
+      - source_labels: [__address__]
+        regex: '(.+):\d+'
+        target_label: __address__
+        replacement: '${1}:8081'
+
   - job_name: 'minio'
     metrics_path: /minio/v2/metrics/cluster
     scheme: http
     nomad_sd_configs:
-      - server: "https://127.0.0.1:4646"
+      - server: "https://{{ nomad_primary_master_address | default(inventory_hostname) }}:4646"
         region: "{{ fact_instance.region }}"
         tls_config:
           insecure_skip_verify: true
         authorization:
-          credentials: "{{ lookup('simple-stack-ui', type='secret', key=inventory_hostname, subkey='nomad_management_token', missing='error') }}"
+          credentials: "{{ lookup('simple-stack-ui', type='secret', key=nomad_primary_master_node | default(inventory_hostname), subkey='nomad_management_token', missing='error') }}"
     relabel_configs:
       - target_label: instance
         replacement: "{{ inventory_hostname }}"
@@ -279,12 +324,12 @@ scrape_configs:
 
   - job_name: 'caddy'
     nomad_sd_configs:
-      - server: "https://127.0.0.1:4646"
+      - server: "https://{{ nomad_primary_master_address | default(inventory_hostname) }}:4646"
         region: "{{ fact_instance.region }}"
         tls_config:
           insecure_skip_verify: true
         authorization:
-          credentials: "{{ lookup('simple-stack-ui', type='secret', key=inventory_hostname, subkey='nomad_management_token', missing='error') }}"
+          credentials: "{{ lookup('simple-stack-ui', type='secret', key=nomad_primary_master_node | default(inventory_hostname), subkey='nomad_management_token', missing='error') }}"
     relabel_configs:
       - target_label: instance
         replacement: "{{ inventory_hostname }}"
@@ -308,14 +353,47 @@ scrape_configs:
       #   replacement: '$1'
       #   target_label: instance
 
+
+  - job_name: 'vllm'
+    nomad_sd_configs:
+      - server: "https://{{ nomad_primary_master_address | default(inventory_hostname) }}:4646"
+        region: "{{ fact_instance.region }}"
+        tls_config:
+          insecure_skip_verify: true
+        authorization:
+          credentials: "{{ lookup('simple-stack-ui', type='secret', key=nomad_primary_master_node | default(inventory_hostname), subkey='nomad_management_token', missing='error') }}"
+    relabel_configs:
+      - target_label: instance
+        replacement: "{{ inventory_hostname }}"
+      - target_label: project
+        replacement: "{{ prometheus_project }}"
+
+      - source_labels: ['__meta_nomad_service']
+        regex: 'vllm'
+        action: keep
+
+      - source_labels: [__meta_nomad_tags]
+        regex: .*,fqdn:([^,]+),.*
+        target_label: __tmp_fqdn
+        replacement: $1
+        action: replace
+
+      - source_labels: [__tmp_fqdn]
+        target_label: fqdn
+
+      # - source_labels: ['__meta_nomad_node']
+      #   replacement: '$1'
+      #   target_label: instance
+
+
   - job_name: 'mysql_exporter'
     nomad_sd_configs:
-      - server: "https://127.0.0.1:4646"
+      - server: "https://{{ nomad_primary_master_address | default(inventory_hostname) }}:4646"
         region: "{{ fact_instance.region }}"
         tls_config:
           insecure_skip_verify: true
         authorization:
-          credentials: "{{ lookup('simple-stack-ui', type='secret', key=inventory_hostname, subkey='nomad_management_token', missing='error') }}"
+          credentials: "{{ lookup('simple-stack-ui', type='secret', key=nomad_primary_master_node | default(inventory_hostname), subkey='nomad_management_token', missing='error') }}"
     relabel_configs:
       - target_label: instance
         replacement: "{{ inventory_hostname }}"
@@ -341,12 +419,12 @@ scrape_configs:
 
   - job_name: 'nginx_exporter'
     nomad_sd_configs:
-      - server: "https://127.0.0.1:4646"
+      - server: "https://{{ nomad_primary_master_address | default(inventory_hostname) }}:4646"
         region: "{{ fact_instance.region }}"
         tls_config:
           insecure_skip_verify: true
         authorization:
-          credentials: "{{ lookup('simple-stack-ui', type='secret', key=inventory_hostname, subkey='nomad_management_token', missing='error') }}"
+          credentials: "{{ lookup('simple-stack-ui', type='secret', key=nomad_primary_master_node | default(inventory_hostname), subkey='nomad_management_token', missing='error') }}"
     relabel_configs:
       - target_label: instance
         replacement: "{{ inventory_hostname }}"
@@ -372,12 +450,12 @@ scrape_configs:
 
   - job_name: 'phpfpm_exporter'
     nomad_sd_configs:
-      - server: "https://127.0.0.1:4646"
+      - server: "https://{{ nomad_primary_master_address | default(inventory_hostname) }}:4646"
         region: "{{ fact_instance.region }}"
         tls_config:
           insecure_skip_verify: true
         authorization:
-          credentials: "{{ lookup('simple-stack-ui', type='secret', key=inventory_hostname, subkey='nomad_management_token', missing='error') }}"
+          credentials: "{{ lookup('simple-stack-ui', type='secret', key=nomad_primary_master_node | default(inventory_hostname), subkey='nomad_management_token', missing='error') }}"
     relabel_configs:
       - target_label: instance
         replacement: "{{ inventory_hostname }}"
@@ -406,12 +484,12 @@ scrape_configs:
     metrics_path: /api/prometheus
     scheme: http
     nomad_sd_configs:
-      - server: "https://127.0.0.1:4646"
+      - server: "https://{{ nomad_primary_master_address | default(inventory_hostname) }}:4646"
         region: "{{ fact_instance.region }}"
         tls_config:
           insecure_skip_verify: true
         authorization:
-          credentials: "{{ lookup('simple-stack-ui', type='secret', key=inventory_hostname, subkey='nomad_management_token', missing='error') }}"
+          credentials: "{{ lookup('simple-stack-ui', type='secret', key=nomad_primary_master_node | default(inventory_hostname), subkey='nomad_management_token', missing='error') }}"
 
     relabel_configs:
       - target_label: instance
diff --git a/ansible/playbooks/paas/roles/promtail/tasks/build.yml b/ansible/playbooks/paas/roles/promtail/tasks/build.yml
index c5b28439..f3b6b120 100644
--- a/ansible/playbooks/paas/roles/promtail/tasks/build.yml
+++ b/ansible/playbooks/paas/roles/promtail/tasks/build.yml
@@ -14,7 +14,7 @@
     url: "{{ upstream_file_url }}"
     dest: "{{ build_work_dir }}/download/"
     mode: '0644'
-    force: no
+    force: false
   register: download_result
 
 - name: Promtail | Unarchive GitHub release
diff --git a/ansible/playbooks/paas/roles/promtail/tasks/main.yml b/ansible/playbooks/paas/roles/promtail/tasks/main.yml
index 69fd20f6..21864a15 100644
--- a/ansible/playbooks/paas/roles/promtail/tasks/main.yml
+++ b/ansible/playbooks/paas/roles/promtail/tasks/main.yml
@@ -1,43 +1,40 @@
 ---
-- name: Promtail | Service is enabled
-  when: loki_remote_write is not defined
-  block:
-    - name: Promtail | Include upstream variables
-      ansible.builtin.include_vars: upstream.yml
+- name: Promtail | Include upstream variables
+  ansible.builtin.include_vars: upstream.yml
 
-    - name: Promtail | Set custom variables
-      ansible.builtin.set_fact:
-        image_version: "{{ latest_version }}"
-        image_name: "{{ image.name }}"
+- name: Promtail | Set custom variables
+  ansible.builtin.set_fact:
+    image_version: "{{ latest_version }}"
+    image_name: "{{ image.name }}"
 
-    - name: Promtail | Get binary
-      include_tasks: build.yml
-      when: ansible_local[image.name] is not defined or ansible_local[image.name] != latest_version
+- name: Promtail | Get binary
+  ansible.builtin.include_tasks: build.yml
+  when: ansible_local[image.name] is not defined or ansible_local[image.name] != latest_version
 
-    - name: Promtail | Create custom directories
-      ansible.builtin.file:
-        dest: "{{ item }}"
-        state: directory
-        recurse: true
-      with_items:
-        - /etc/promtail
-        - /var/lib/promtail
+- name: Promtail | Create custom directories
+  ansible.builtin.file:
+    dest: "{{ item }}"
+    state: directory
+    recurse: true
+  with_items:
+    - /etc/promtail
+    - /var/lib/promtail
 
-    - name: Promtail | Copy templates
-      ansible.builtin.template:
-        src: "{{ item.src }}"
-        dest: "{{ item.dest }}"
-        mode: 0644
-        owner: root
-        group: root
-      loop:
-        - src: default.j2
-          dest: /etc/default/promtail
-        - src: config.yaml.j2
-          dest: /etc/promtail/config.yaml
-        - src: service.j2
-          dest: /etc/systemd/system/promtail.service
-      notify: Restart promtail
+- name: Promtail | Copy templates
+  ansible.builtin.template:
+    src: "{{ item.src }}"
+    dest: "{{ item.dest }}"
+    mode: '0644'
+    owner: root
+    group: root
+  loop:
+    - src: default.j2
+      dest: /etc/default/promtail
+    - src: config.yaml.j2
+      dest: /etc/promtail/config.yaml
+    - src: service.j2
+      dest: /etc/systemd/system/promtail.service
+  notify: Restart promtail
 
-    - name: Promtail | Flush handlers
-      ansible.builtin.meta: flush_handlers
+- name: Promtail | Flush handlers
+  ansible.builtin.meta: flush_handlers
diff --git a/ansible/playbooks/paas/roles/promtail/templates/config.yaml.j2 b/ansible/playbooks/paas/roles/promtail/templates/config.yaml.j2
index ac2220d3..50a2def9 100644
--- a/ansible/playbooks/paas/roles/promtail/templates/config.yaml.j2
+++ b/ansible/playbooks/paas/roles/promtail/templates/config.yaml.j2
@@ -1,5 +1,5 @@
 server:
-  http_listen_address: 127.0.0.1
+  http_listen_address: {{ hostvars[inventory_hostname]['ansible_' + nomad_iface]['ipv4']['address'] }}
   http_listen_port: 9080
   grpc_listen_port: 0
   log_level: warn
@@ -7,11 +7,17 @@ server:
 positions:
   filename: /var/lib/promtail/positions.yaml
 
+{% if loki_remote_write is defined %}
 clients:
   - url: {{ loki_remote_write.url }}/api/prom/push
+{% if loki_remote_write.basic_auth %}
     basic_auth:
       username: {{ loki_remote_write.login }}
       password: {{ loki_remote_write.password }}
+{% endif %}
+{% else %}
+clients: []
+{% endif %}
 
 scrape_configs:
   - job_name: system
diff --git a/ansible/playbooks/paas/roles/systemd_exporter/templates/default.j2 b/ansible/playbooks/paas/roles/systemd_exporter/templates/default.j2
index 1a27d383..f64efe14 100644
--- a/ansible/playbooks/paas/roles/systemd_exporter/templates/default.j2
+++ b/ansible/playbooks/paas/roles/systemd_exporter/templates/default.j2
@@ -1,2 +1,2 @@
-ARGS="--web.listen-address=127.0.0.1:9558 \
+ARGS="--web.listen-address={{ hostvars[inventory_hostname]['ansible_' + nomad_iface]['ipv4']['address'] }}:9558 \
 --systemd.collector.unit-include=docker.service|promtail.service|coredns.service|prometheus.service|blackbox_exporter.service|node_exporter.service|scan_exporter.service"
diff --git a/ansible/playbooks/paas/systemd-resolved.yml b/ansible/playbooks/paas/systemd-resolved.yml
index f07ac0bb..9cfc4f6d 100644
--- a/ansible/playbooks/paas/systemd-resolved.yml
+++ b/ansible/playbooks/paas/systemd-resolved.yml
@@ -16,6 +16,7 @@
         content: |
           [Resolve]
           DNSStubListener=yes
+          DNSStubListenerExtra=172.17.0.1:53
         dest: /etc/systemd/resolved.conf.d/systemd-resolved.conf
         mode: '0644'
       notify: Restart systemd-resolved
diff --git a/ansible/playbooks/saas/basic_auth.yml b/ansible/playbooks/saas/basic_auth.yml
index 06e2d7af..cb400913 100644
--- a/ansible/playbooks/saas/basic_auth.yml
+++ b/ansible/playbooks/saas/basic_auth.yml
@@ -14,4 +14,4 @@
   tasks:
     - name: Debug hash
       ansible.builtin.debug:
-        msg: "{{ login }}:{{ password | password_hash('blowfish') }}"
+        msg: "{{ login }}:{{ password | ansible.builtin.password_hash('blowfish') }}"
diff --git a/ansible/playbooks/saas/image.yml b/ansible/playbooks/saas/image.yml
index 9f819d3d..63b71ebd 100644
--- a/ansible/playbooks/saas/image.yml
+++ b/ansible/playbooks/saas/image.yml
@@ -26,25 +26,9 @@
         state: directory
         mode: '0755'
       loop:
-        - /root/.docker
         - "{{ build_work_dir }}/download"
         - "{{ build_work_dir }}/{{ upstream_default_arch }}"
 
-    - name: Copy docker config file
-      ansible.builtin.copy:
-        content: |
-          {
-              "auths": {
-                  "{{ docker_private_registry.url }}": {
-                      "auth": "{{ (docker_private_registry.username + ':' + docker_private_registry.password) | b64encode }}"
-                  }
-              }
-          }
-        dest: /root/.docker/config.json
-        owner: root
-        group: root
-        mode: '0600'
-
   tasks:
     - name: Install dependencies
       ansible.builtin.include_role:
@@ -59,9 +43,9 @@
     - name: Build
       when: image_build
       block:
-        - name: Build image
+        - name: Build and publish image
           community.docker.docker_image_build:
-            name: "{{ docker_private_registry.url }}/{{ image_name }}:{{ image_version }}"
+            name: "{{ docker_private_registry.url }}/{{ docker_private_registry.project is defined | ternary(docker_private_registry.project + '/', '') }}{{ image_name }}:{{ image_version }}"
             tag: latest
             path: "/tmp/{{ catalog }}"
             dockerfile: Dockerfile
@@ -91,6 +75,7 @@
             version: "{{ image_version }}"
         force_basic_auth: true
         status_code: 200
-      ignore_errors: true
+      register: ui_update
+      failed_when: ui_update.status != 200
       delegate_to: localhost
-      become: false
\ No newline at end of file
+      become: false
diff --git a/ansible/playbooks/saas/main.yml b/ansible/playbooks/saas/main.yml
index 0097343b..07729907 100644
--- a/ansible/playbooks/saas/main.yml
+++ b/ansible/playbooks/saas/main.yml
@@ -30,8 +30,8 @@
         software: "{{ lookup('simple-stack-ui', type='software', key=domain, subkey='', missing='warn') }}"
 
     - name: Debug software
-      debug:
-        msg: "{{ software }}" 
+      ansible.builtin.debug:
+        msg: "{{ software }}"
 
   tasks:
     - name: Deploy service
diff --git a/ansible/playbooks/saas/roles/adguard/templates/nomad.hcl b/ansible/playbooks/saas/roles/adguard/templates/nomad.hcl
index c3a315a5..3bb996f3 100644
--- a/ansible/playbooks/saas/roles/adguard/templates/nomad.hcl
+++ b/ansible/playbooks/saas/roles/adguard/templates/nomad.hcl
@@ -3,7 +3,7 @@ job "{{ domain }}" {
   datacenters = ["{{ fact_instance.datacenter }}"]
   type = "service"
 
-{% if software.constraints.location %}
+{% if software.constraints is defined and software.constraints.location is defined %}
   constraint {
     attribute    = "${meta.location}"
     set_contains = "{{ software.constraints.location }}"
diff --git a/ansible/playbooks/saas/roles/arangodb/templates/nomad.hcl b/ansible/playbooks/saas/roles/arangodb/templates/nomad.hcl
index e6956e43..ccd7cace 100644
--- a/ansible/playbooks/saas/roles/arangodb/templates/nomad.hcl
+++ b/ansible/playbooks/saas/roles/arangodb/templates/nomad.hcl
@@ -3,7 +3,7 @@ job "{{ domain }}" {
   datacenters = ["{{ fact_instance.datacenter }}"]
   type = "service"
 
-{% if software.constraints.location %}
+{% if software.constraints is defined and software.constraints.location is defined %}
   constraint {
     attribute    = "${meta.location}"
     set_contains = "{{ software.constraints.location }}"
diff --git a/ansible/playbooks/saas/roles/caddy/tasks/main.yml b/ansible/playbooks/saas/roles/caddy/tasks/main.yml
index b45d0b3e..f61aa1fa 100644
--- a/ansible/playbooks/saas/roles/caddy/tasks/main.yml
+++ b/ansible/playbooks/saas/roles/caddy/tasks/main.yml
@@ -1,15 +1,4 @@
 ---
-- name: Create default directory
-  ansible.builtin.file:
-    path: "{{ item }}"
-    state: directory
-    owner: root
-    group: root
-    mode: "0755"
-  loop:
-    - "{{ software_path }}/etc/caddy"
-  delegate_to: "{{ software.instance }}"
-
 - name: Copy nomad job to destination
   ansible.builtin.template:
     src: nomad.hcl
diff --git a/ansible/playbooks/saas/roles/caddy/templates/Dockerfile.j2 b/ansible/playbooks/saas/roles/caddy/templates/Dockerfile.j2
index 387b89ba..a423587b 100644
--- a/ansible/playbooks/saas/roles/caddy/templates/Dockerfile.j2
+++ b/ansible/playbooks/saas/roles/caddy/templates/Dockerfile.j2
@@ -15,4 +15,4 @@ RUN mkdir -p /var/log/caddy /var/lib/caddy /etc/caddy \
 
 USER caddy
 
-CMD ["caddy", "run", "--config", "/etc/caddy/Caddyfile"]
+CMD ["caddy", "run"]
diff --git a/ansible/playbooks/saas/roles/caddy/templates/nomad.hcl b/ansible/playbooks/saas/roles/caddy/templates/nomad.hcl
index edbe32a0..cb828014 100644
--- a/ansible/playbooks/saas/roles/caddy/templates/nomad.hcl
+++ b/ansible/playbooks/saas/roles/caddy/templates/nomad.hcl
@@ -3,17 +3,19 @@ job "{{ domain }}" {
   datacenters = ["{{ fact_instance.datacenter }}"]
   type = "service"
 
-{% if software.constraints.location %}
+{% if software.constraints is defined and software.constraints.location is defined %}
   constraint {
     attribute    = "${meta.location}"
     set_contains = "{{ software.constraints.location }}"
   }
 {% endif %}
 
+{% if software.constraints is defined and software.constraints.instance is defined %}
   constraint {
     attribute    = "${meta.instance}"
-    set_contains = "{{ software.instance }}"
+    set_contains = "{{ software.constraints.instance }}"
   }
+{% endif %}
 
   group "{{ domain }}" {
     count = {{ software.scale | default(1) }}
@@ -49,22 +51,26 @@ job "{{ domain }}" {
     task "{{ domain }}-caddy" {
       driver = "docker"
 
-      config {
-        image = "{{ docker_private_registry.url }}/caddy:{{ softwares.caddy.version }}"
-        volumes = [
-          "{{ software_path }}/etc/caddy:/etc/caddy:ro"
-        ]
-        ports = ["caddy", "metrics"]
-      }
-
       template {
-        change_mode = "noop"
-        destination = "{{ software_path }}/etc/caddy/Caddyfile"
+        change_mode = "restart"
+        destination = "local/Caddyfile"
+        perms = "644"
         data = <<EOH
 {{ caddy_config }}
   EOH
       }
 
+      config {
+        image = "{{ docker_private_registry.url }}/caddy:{{ softwares.caddy.version }}"
+        ports = ["caddy", "metrics"]
+        command = "caddy"
+        args  = [
+          "run",
+          "--config", 
+          "/local/Caddyfile"
+        ]
+      }
+
       resources {
         cpu    = {{ size[software.size].cpu }}
         memory = {{ size[software.size].memory }}
diff --git a/ansible/playbooks/saas/roles/dolibarr/templates/nomad.hcl b/ansible/playbooks/saas/roles/dolibarr/templates/nomad.hcl
index a0faff30..a3c77c12 100644
--- a/ansible/playbooks/saas/roles/dolibarr/templates/nomad.hcl
+++ b/ansible/playbooks/saas/roles/dolibarr/templates/nomad.hcl
@@ -3,7 +3,7 @@ job "{{ domain }}" {
   datacenters = ["{{ fact_instance.datacenter }}"]
   type = "service"
 
-{% if software.constraints.location %}
+{% if software.constraints is defined and software.constraints.location is defined %}
   constraint {
     attribute    = "${meta.location}"
     set_contains = "{{ software.constraints.location }}"
diff --git a/ansible/playbooks/saas/roles/forgejo/templates/nomad.hcl b/ansible/playbooks/saas/roles/forgejo/templates/nomad.hcl
index 1608d244..0a93e435 100644
--- a/ansible/playbooks/saas/roles/forgejo/templates/nomad.hcl
+++ b/ansible/playbooks/saas/roles/forgejo/templates/nomad.hcl
@@ -3,7 +3,7 @@ job "{{ domain }}" {
   datacenters = ["{{ fact_instance.datacenter }}"]
   type = "service"
 
-{% if software.constraints.location %}
+{% if software.constraints is defined and software.constraints.location is defined %}
   constraint {
     attribute    = "${meta.location}"
     set_contains = "{{ software.constraints.location }}"
diff --git a/ansible/playbooks/saas/roles/freshrss/templates/nomad.hcl b/ansible/playbooks/saas/roles/freshrss/templates/nomad.hcl
index 5b924ce5..466b31ea 100644
--- a/ansible/playbooks/saas/roles/freshrss/templates/nomad.hcl
+++ b/ansible/playbooks/saas/roles/freshrss/templates/nomad.hcl
@@ -3,7 +3,7 @@ job "{{ domain }}" {
   datacenters = ["{{ fact_instance.datacenter }}"]
   type = "service"
 
-{% if software.constraints.location %}
+{% if software.constraints is defined and software.constraints.location is defined %}
   constraint {
     attribute    = "${meta.location}"
     set_contains = "{{ software.constraints.location }}"
diff --git a/ansible/playbooks/saas/roles/grafana/files/dashboards/nomad/nomad.json b/ansible/playbooks/saas/roles/grafana/files/dashboards/nomad/nomad.json
index 0a22af36..2c2798fe 100644
--- a/ansible/playbooks/saas/roles/grafana/files/dashboards/nomad/nomad.json
+++ b/ansible/playbooks/saas/roles/grafana/files/dashboards/nomad/nomad.json
@@ -18,11 +18,9 @@
   "description": "Nomad Jobs metrics",
   "editable": true,
   "fiscalYearStartMonth": 0,
-  "gnetId": 6281,
   "graphTooltip": 0,
-  "id": 35,
+  "id": 45,
   "links": [],
-  "liveNow": false,
   "panels": [
     {
       "datasource": {
@@ -41,6 +39,7 @@
             "axisLabel": "",
             "axisPlacement": "auto",
             "barAlignment": 0,
+            "barWidthFactor": 0.6,
             "drawStyle": "line",
             "fillOpacity": 0,
             "gradientMode": "none",
@@ -57,6 +56,7 @@
               "type": "linear"
             },
             "showPoints": "auto",
+            "showValues": false,
             "spanNulls": false,
             "stacking": {
               "group": "A",
@@ -72,7 +72,7 @@
             "steps": [
               {
                 "color": "green",
-                "value": null
+                "value": 0
               },
               {
                 "color": "red",
@@ -100,11 +100,12 @@
           "showLegend": true
         },
         "tooltip": {
+          "hideZeros": false,
           "mode": "single",
           "sort": "none"
         }
       },
-      "pluginVersion": "10.2.2",
+      "pluginVersion": "12.2.0",
       "targets": [
         {
           "datasource": {
@@ -112,7 +113,7 @@
             "uid": "prometheus"
           },
           "editorMode": "code",
-          "expr": "nomad_client_allocs_failed{host=~\"$host\"} ",
+          "expr": "nomad_client_allocs_failed{instance=~\"$instance\"} ",
           "instant": false,
           "legendFormat": "{{exported_job}}",
           "range": true,
@@ -139,6 +140,7 @@
             "axisLabel": "",
             "axisPlacement": "auto",
             "barAlignment": 0,
+            "barWidthFactor": 0.6,
             "drawStyle": "line",
             "fillOpacity": 0,
             "gradientMode": "none",
@@ -155,6 +157,7 @@
               "type": "linear"
             },
             "showPoints": "auto",
+            "showValues": false,
             "spanNulls": false,
             "stacking": {
               "group": "A",
@@ -170,7 +173,7 @@
             "steps": [
               {
                 "color": "green",
-                "value": null
+                "value": 0
               },
               {
                 "color": "red",
@@ -198,11 +201,12 @@
           "showLegend": true
         },
         "tooltip": {
+          "hideZeros": false,
           "mode": "single",
           "sort": "none"
         }
       },
-      "pluginVersion": "10.2.2",
+      "pluginVersion": "12.2.0",
       "targets": [
         {
           "datasource": {
@@ -210,7 +214,7 @@
             "uid": "prometheus"
           },
           "editorMode": "code",
-          "expr": "sum(nomad_client_allocations_pending{host=~\"$host\",host=~\"$host\"}) by(project, node_status)",
+          "expr": "sum(nomad_client_allocations_pending{instance=~\"$instance\"}) by(project, node_status)",
           "instant": false,
           "legendFormat": "{{project}} {{node_status}}",
           "range": true,
@@ -250,6 +254,7 @@
             "axisLabel": "",
             "axisPlacement": "auto",
             "barAlignment": 0,
+            "barWidthFactor": 0.6,
             "drawStyle": "line",
             "fillOpacity": 10,
             "gradientMode": "none",
@@ -266,6 +271,7 @@
               "type": "linear"
             },
             "showPoints": "never",
+            "showValues": false,
             "spanNulls": false,
             "stacking": {
               "group": "A",
@@ -282,7 +288,7 @@
             "steps": [
               {
                 "color": "green",
-                "value": null
+                "value": 0
               },
               {
                 "color": "red",
@@ -301,7 +307,6 @@
         "y": 9
       },
       "id": 2,
-      "links": [],
       "options": {
         "legend": {
           "calcs": [
@@ -311,14 +316,17 @@
           ],
           "displayMode": "table",
           "placement": "right",
-          "showLegend": true
+          "showLegend": true,
+          "sortBy": "Max",
+          "sortDesc": true
         },
         "tooltip": {
+          "hideZeros": false,
           "mode": "multi",
           "sort": "none"
         }
       },
-      "pluginVersion": "10.2.2",
+      "pluginVersion": "12.2.0",
       "repeat": "host",
       "repeatDirection": "v",
       "targets": [
@@ -327,134 +335,140 @@
             "datasource": "$datasource",
             "uid": "$datasource"
           },
-          "expr": "avg(nomad_client_allocs_cpu_total_percent{host=~\"$host\"}) by(exported_job, task)",
+          "editorMode": "code",
+          "expr": "avg(nomad_client_allocs_cpu_total_percent{instance=~\"$instance\"}) by(exported_job, task)",
           "format": "time_series",
           "interval": "",
           "intervalFactor": 1,
           "legendFormat": "{{task}}",
+          "range": true,
           "refId": "A"
         }
       ],
-      "title": "CPU Usage Percent from $project / $host",
+      "title": "CPU Usage Percent from $project / $instance",
       "type": "timeseries"
     },
     {
-      "collapsed": true,
+      "collapsed": false,
       "gridPos": {
         "h": 1,
         "w": 24,
         "x": 0,
-        "y": 90
+        "y": 18
       },
       "id": 9,
-      "panels": [
-        {
-          "datasource": {
-            "datasource": "$datasource",
-            "uid": "$datasource"
+      "panels": [],
+      "title": "CPU total ticks",
+      "type": "row"
+    },
+    {
+      "datasource": {
+        "datasource": "$datasource",
+        "uid": "$datasource"
+      },
+      "fieldConfig": {
+        "defaults": {
+          "color": {
+            "mode": "palette-classic"
           },
-          "fieldConfig": {
-            "defaults": {
-              "color": {
-                "mode": "palette-classic"
-              },
-              "custom": {
-                "axisBorderShow": false,
-                "axisCenteredZero": false,
-                "axisColorMode": "text",
-                "axisLabel": "",
-                "axisPlacement": "auto",
-                "barAlignment": 0,
-                "drawStyle": "line",
-                "fillOpacity": 10,
-                "gradientMode": "none",
-                "hideFrom": {
-                  "legend": false,
-                  "tooltip": false,
-                  "viz": false
-                },
-                "insertNulls": false,
-                "lineInterpolation": "linear",
-                "lineWidth": 1,
-                "pointSize": 5,
-                "scaleDistribution": {
-                  "type": "linear"
-                },
-                "showPoints": "never",
-                "spanNulls": false,
-                "stacking": {
-                  "group": "A",
-                  "mode": "none"
-                },
-                "thresholdsStyle": {
-                  "mode": "off"
-                }
-              },
-              "decimals": 3,
-              "mappings": [],
-              "thresholds": {
-                "mode": "absolute",
-                "steps": [
-                  {
-                    "color": "green"
-                  },
-                  {
-                    "color": "red",
-                    "value": 80
-                  }
-                ]
-              },
-              "unit": "timeticks"
+          "custom": {
+            "axisBorderShow": false,
+            "axisCenteredZero": false,
+            "axisColorMode": "text",
+            "axisLabel": "",
+            "axisPlacement": "auto",
+            "barAlignment": 0,
+            "barWidthFactor": 0.6,
+            "drawStyle": "line",
+            "fillOpacity": 10,
+            "gradientMode": "none",
+            "hideFrom": {
+              "legend": false,
+              "tooltip": false,
+              "viz": false
             },
-            "overrides": []
-          },
-          "gridPos": {
-            "h": 8,
-            "w": 24,
-            "x": 0,
-            "y": 10
-          },
-          "id": 3,
-          "links": [],
-          "options": {
-            "legend": {
-              "calcs": [
-                "lastNotNull",
-                "max",
-                "min"
-              ],
-              "displayMode": "table",
-              "placement": "right",
-              "showLegend": true
+            "insertNulls": false,
+            "lineInterpolation": "linear",
+            "lineWidth": 1,
+            "pointSize": 5,
+            "scaleDistribution": {
+              "type": "linear"
+            },
+            "showPoints": "never",
+            "showValues": false,
+            "spanNulls": false,
+            "stacking": {
+              "group": "A",
+              "mode": "none"
             },
-            "tooltip": {
-              "mode": "multi",
-              "sort": "none"
+            "thresholdsStyle": {
+              "mode": "off"
             }
           },
-          "pluginVersion": "10.2.2",
-          "repeat": "host",
-          "repeatDirection": "v",
-          "targets": [
-            {
-              "datasource": {
-                "datasource": "$datasource",
-                "uid": "$datasource"
+          "decimals": 3,
+          "mappings": [],
+          "thresholds": {
+            "mode": "absolute",
+            "steps": [
+              {
+                "color": "green",
+                "value": 0
               },
-              "expr": "avg(nomad_client_allocs_cpu_total_ticks{host=~\"$host\"}) by(exported_job, task)",
-              "format": "time_series",
-              "interval": "",
-              "intervalFactor": 1,
-              "legendFormat": "{{task}}",
-              "refId": "A"
-            }
+              {
+                "color": "red",
+                "value": 80
+              }
+            ]
+          },
+          "unit": "timeticks"
+        },
+        "overrides": []
+      },
+      "gridPos": {
+        "h": 8,
+        "w": 24,
+        "x": 0,
+        "y": 19
+      },
+      "id": 3,
+      "options": {
+        "legend": {
+          "calcs": [
+            "lastNotNull",
+            "max",
+            "min"
           ],
-          "title": "CPU Total Ticks from $project / $host",
-          "type": "timeseries"
+          "displayMode": "table",
+          "placement": "right",
+          "showLegend": true
+        },
+        "tooltip": {
+          "hideZeros": false,
+          "mode": "multi",
+          "sort": "none"
+        }
+      },
+      "pluginVersion": "12.2.0",
+      "repeat": "host",
+      "repeatDirection": "v",
+      "targets": [
+        {
+          "datasource": {
+            "datasource": "$datasource",
+            "uid": "$datasource"
+          },
+          "editorMode": "code",
+          "expr": "avg(nomad_client_allocs_cpu_total_ticks{instance=~\"$instance\"}) by(exported_job, task)",
+          "format": "time_series",
+          "interval": "",
+          "intervalFactor": 1,
+          "legendFormat": "{{task}}",
+          "range": true,
+          "refId": "A"
         }
       ],
-      "title": "CPU total ticks",
-      "type": "row"
+      "title": "CPU Total Ticks from $project / $instance",
+      "type": "timeseries"
     },
     {
       "collapsed": false,
@@ -462,7 +476,7 @@
         "h": 1,
         "w": 24,
         "x": 0,
-        "y": 91
+        "y": 27
       },
       "id": 13,
       "panels": [],
@@ -486,6 +500,7 @@
             "axisLabel": "",
             "axisPlacement": "auto",
             "barAlignment": 0,
+            "barWidthFactor": 0.6,
             "drawStyle": "line",
             "fillOpacity": 10,
             "gradientMode": "none",
@@ -502,6 +517,7 @@
               "type": "linear"
             },
             "showPoints": "never",
+            "showValues": false,
             "spanNulls": false,
             "stacking": {
               "group": "A",
@@ -517,7 +533,7 @@
             "steps": [
               {
                 "color": "green",
-                "value": null
+                "value": 0
               },
               {
                 "color": "red",
@@ -533,11 +549,10 @@
         "h": 7,
         "w": 24,
         "x": 0,
-        "y": 92
+        "y": 28
       },
       "id": 12,
       "interval": "1m",
-      "links": [],
       "options": {
         "legend": {
           "calcs": [
@@ -550,11 +565,12 @@
           "sortDesc": true
         },
         "tooltip": {
+          "hideZeros": false,
           "mode": "multi",
           "sort": "none"
         }
       },
-      "pluginVersion": "10.2.2",
+      "pluginVersion": "12.2.0",
       "repeat": "host",
       "repeatDirection": "v",
       "targets": [
@@ -564,7 +580,7 @@
             "uid": "$datasource"
           },
           "editorMode": "code",
-          "expr": "avg(nomad_client_allocs_memory_usage{host=~\"$host\"}) by(exported_job, task)",
+          "expr": "avg(nomad_client_allocs_memory_usage{instance=~\"$instance\"}) by(exported_job, task)",
           "format": "time_series",
           "interval": "",
           "intervalFactor": 1,
@@ -573,275 +589,276 @@
           "refId": "A"
         }
       ],
-      "title": "Memory usage from $project / $host",
+      "title": "Memory usage from $project / $instance",
       "type": "timeseries"
     },
     {
-      "collapsed": true,
+      "collapsed": false,
       "gridPos": {
         "h": 1,
         "w": 24,
         "x": 0,
-        "y": 155
+        "y": 35
       },
       "id": 11,
-      "panels": [
-        {
-          "datasource": {
-            "datasource": "$datasource",
-            "uid": "$datasource"
+      "panels": [],
+      "title": "Memory cache",
+      "type": "row"
+    },
+    {
+      "datasource": {
+        "datasource": "$datasource",
+        "uid": "$datasource"
+      },
+      "fieldConfig": {
+        "defaults": {
+          "color": {
+            "mode": "palette-classic"
           },
-          "fieldConfig": {
-            "defaults": {
-              "color": {
-                "mode": "palette-classic"
-              },
-              "custom": {
-                "axisBorderShow": false,
-                "axisCenteredZero": false,
-                "axisColorMode": "text",
-                "axisLabel": "",
-                "axisPlacement": "auto",
-                "barAlignment": 0,
-                "drawStyle": "line",
-                "fillOpacity": 10,
-                "gradientMode": "none",
-                "hideFrom": {
-                  "legend": false,
-                  "tooltip": false,
-                  "viz": false
-                },
-                "insertNulls": false,
-                "lineInterpolation": "linear",
-                "lineWidth": 1,
-                "pointSize": 5,
-                "scaleDistribution": {
-                  "type": "linear"
-                },
-                "showPoints": "never",
-                "spanNulls": false,
-                "stacking": {
-                  "group": "A",
-                  "mode": "none"
-                },
-                "thresholdsStyle": {
-                  "mode": "off"
-                }
-              },
-              "decimals": 3,
-              "mappings": [],
-              "thresholds": {
-                "mode": "absolute",
-                "steps": [
-                  {
-                    "color": "green"
-                  },
-                  {
-                    "color": "red",
-                    "value": 80
-                  }
-                ]
-              },
-              "unit": "decbytes"
+          "custom": {
+            "axisBorderShow": false,
+            "axisCenteredZero": false,
+            "axisColorMode": "text",
+            "axisLabel": "",
+            "axisPlacement": "auto",
+            "barAlignment": 0,
+            "barWidthFactor": 0.6,
+            "drawStyle": "line",
+            "fillOpacity": 10,
+            "gradientMode": "none",
+            "hideFrom": {
+              "legend": false,
+              "tooltip": false,
+              "viz": false
             },
-            "overrides": []
-          },
-          "gridPos": {
-            "h": 7,
-            "w": 24,
-            "x": 0,
-            "y": 12
-          },
-          "id": 7,
-          "links": [],
-          "options": {
-            "legend": {
-              "calcs": [
-                "lastNotNull",
-                "max",
-                "min"
-              ],
-              "displayMode": "table",
-              "placement": "right",
-              "showLegend": true
+            "insertNulls": false,
+            "lineInterpolation": "linear",
+            "lineWidth": 1,
+            "pointSize": 5,
+            "scaleDistribution": {
+              "type": "linear"
+            },
+            "showPoints": "never",
+            "showValues": false,
+            "spanNulls": false,
+            "stacking": {
+              "group": "A",
+              "mode": "none"
             },
-            "tooltip": {
-              "mode": "multi",
-              "sort": "none"
+            "thresholdsStyle": {
+              "mode": "off"
             }
           },
-          "pluginVersion": "10.2.2",
-          "repeat": "host",
-          "repeatDirection": "v",
-          "targets": [
-            {
-              "datasource": {
-                "datasource": "$datasource",
-                "uid": "$datasource"
+          "decimals": 3,
+          "mappings": [],
+          "thresholds": {
+            "mode": "absolute",
+            "steps": [
+              {
+                "color": "green",
+                "value": 0
               },
-              "expr": "avg(nomad_client_allocs_memory_cache{host=~\"$host\"}) by(exported_job, task)",
-              "format": "time_series",
-              "interval": "",
-              "intervalFactor": 1,
-              "legendFormat": "{{task}}",
-              "refId": "A"
-            }
+              {
+                "color": "red",
+                "value": 80
+              }
+            ]
+          },
+          "unit": "decbytes"
+        },
+        "overrides": []
+      },
+      "gridPos": {
+        "h": 7,
+        "w": 24,
+        "x": 0,
+        "y": 36
+      },
+      "id": 7,
+      "options": {
+        "legend": {
+          "calcs": [
+            "lastNotNull",
+            "max",
+            "min"
           ],
-          "title": "Memory Cache from $project / $host",
-          "type": "timeseries"
+          "displayMode": "table",
+          "placement": "right",
+          "showLegend": true
+        },
+        "tooltip": {
+          "hideZeros": false,
+          "mode": "multi",
+          "sort": "none"
+        }
+      },
+      "pluginVersion": "12.2.0",
+      "repeat": "host",
+      "repeatDirection": "v",
+      "targets": [
+        {
+          "datasource": {
+            "datasource": "$datasource",
+            "uid": "$datasource"
+          },
+          "editorMode": "code",
+          "expr": "avg(nomad_client_allocs_memory_cache{instance=~\"$instance\"}) by(exported_job, task)",
+          "format": "time_series",
+          "interval": "",
+          "intervalFactor": 1,
+          "legendFormat": "{{task}}",
+          "range": true,
+          "refId": "A"
         }
       ],
-      "title": "Memory cache",
-      "type": "row"
+      "title": "Memory Cache from $project / $instance",
+      "type": "timeseries"
     },
     {
-      "collapsed": true,
+      "collapsed": false,
       "gridPos": {
         "h": 1,
         "w": 24,
         "x": 0,
-        "y": 156
+        "y": 43
       },
       "id": 10,
-      "panels": [
-        {
-          "datasource": {
-            "datasource": "$datasource",
-            "uid": "$datasource"
+      "panels": [],
+      "title": "RSS",
+      "type": "row"
+    },
+    {
+      "datasource": {
+        "datasource": "$datasource",
+        "uid": "$datasource"
+      },
+      "fieldConfig": {
+        "defaults": {
+          "color": {
+            "mode": "palette-classic"
           },
-          "fieldConfig": {
-            "defaults": {
-              "color": {
-                "mode": "palette-classic"
-              },
-              "custom": {
-                "axisBorderShow": false,
-                "axisCenteredZero": false,
-                "axisColorMode": "text",
-                "axisLabel": "",
-                "axisPlacement": "auto",
-                "barAlignment": 0,
-                "drawStyle": "line",
-                "fillOpacity": 10,
-                "gradientMode": "none",
-                "hideFrom": {
-                  "legend": false,
-                  "tooltip": false,
-                  "viz": false
-                },
-                "insertNulls": false,
-                "lineInterpolation": "linear",
-                "lineWidth": 1,
-                "pointSize": 5,
-                "scaleDistribution": {
-                  "type": "linear"
-                },
-                "showPoints": "never",
-                "spanNulls": false,
-                "stacking": {
-                  "group": "A",
-                  "mode": "none"
-                },
-                "thresholdsStyle": {
-                  "mode": "off"
-                }
-              },
-              "decimals": 3,
-              "mappings": [],
-              "thresholds": {
-                "mode": "absolute",
-                "steps": [
-                  {
-                    "color": "green"
-                  },
-                  {
-                    "color": "red",
-                    "value": 80
-                  }
-                ]
-              },
-              "unit": "decbytes"
+          "custom": {
+            "axisBorderShow": false,
+            "axisCenteredZero": false,
+            "axisColorMode": "text",
+            "axisLabel": "",
+            "axisPlacement": "auto",
+            "barAlignment": 0,
+            "barWidthFactor": 0.6,
+            "drawStyle": "line",
+            "fillOpacity": 10,
+            "gradientMode": "none",
+            "hideFrom": {
+              "legend": false,
+              "tooltip": false,
+              "viz": false
             },
-            "overrides": []
-          },
-          "gridPos": {
-            "h": 8,
-            "w": 24,
-            "x": 0,
-            "y": 13
-          },
-          "id": 6,
-          "links": [],
-          "options": {
-            "legend": {
-              "calcs": [
-                "lastNotNull",
-                "max",
-                "min"
-              ],
-              "displayMode": "table",
-              "placement": "right",
-              "showLegend": true
+            "insertNulls": false,
+            "lineInterpolation": "linear",
+            "lineWidth": 1,
+            "pointSize": 5,
+            "scaleDistribution": {
+              "type": "linear"
+            },
+            "showPoints": "never",
+            "showValues": false,
+            "spanNulls": false,
+            "stacking": {
+              "group": "A",
+              "mode": "none"
             },
-            "tooltip": {
-              "mode": "multi",
-              "sort": "none"
+            "thresholdsStyle": {
+              "mode": "off"
             }
           },
-          "pluginVersion": "10.2.2",
-          "repeat": "host",
-          "repeatDirection": "v",
-          "targets": [
-            {
-              "datasource": {
-                "datasource": "$datasource",
-                "uid": "$datasource"
+          "decimals": 3,
+          "mappings": [],
+          "thresholds": {
+            "mode": "absolute",
+            "steps": [
+              {
+                "color": "green",
+                "value": 0
               },
-              "expr": "avg(nomad_client_allocs_memory_rss{host=~\"$host\"}) by(exported_job, task)",
-              "format": "time_series",
-              "interval": "",
-              "intervalFactor": 1,
-              "legendFormat": "{{task}}",
-              "refId": "A"
-            }
+              {
+                "color": "red",
+                "value": 80
+              }
+            ]
+          },
+          "unit": "decbytes"
+        },
+        "overrides": []
+      },
+      "gridPos": {
+        "h": 8,
+        "w": 24,
+        "x": 0,
+        "y": 44
+      },
+      "id": 6,
+      "options": {
+        "legend": {
+          "calcs": [
+            "lastNotNull",
+            "max",
+            "min"
           ],
-          "title": "RSS from $project / $host",
-          "type": "timeseries"
+          "displayMode": "table",
+          "placement": "right",
+          "showLegend": true
+        },
+        "tooltip": {
+          "hideZeros": false,
+          "mode": "multi",
+          "sort": "none"
+        }
+      },
+      "pluginVersion": "12.2.0",
+      "repeat": "host",
+      "repeatDirection": "v",
+      "targets": [
+        {
+          "datasource": {
+            "datasource": "$datasource",
+            "uid": "$datasource"
+          },
+          "editorMode": "code",
+          "expr": "avg(nomad_client_allocs_memory_rss{instance=~\"$instance\"}) by(exported_job, task)",
+          "format": "time_series",
+          "interval": "",
+          "intervalFactor": 1,
+          "legendFormat": "{{task}}",
+          "range": true,
+          "refId": "A"
         }
       ],
-      "title": "RSS",
-      "type": "row"
+      "title": "RSS from $project / $instance",
+      "type": "timeseries"
     }
   ],
+  "preload": false,
   "refresh": "30s",
-  "schemaVersion": 38,
+  "schemaVersion": 42,
   "tags": [],
   "templating": {
     "list": [
       {
         "current": {
-          "selected": false,
-          "text": "Mimir",
-          "value": "prometheus"
+          "text": "prometheus",
+          "value": "cf0jbbq5bcwsgb"
         },
-        "hide": 0,
         "includeAll": false,
-        "multi": false,
         "name": "datasource",
         "options": [],
         "query": "prometheus",
-        "queryValue": "",
         "refresh": 1,
         "regex": "",
-        "skipUrlSync": false,
         "type": "datasource"
       },
       {
         "current": {
-          "selected": true,
-          "text": [
-            "All"
-          ],
+          "text": "All",
           "value": [
             "$__all"
           ]
@@ -851,7 +868,6 @@
           "uid": "${datasource}"
         },
         "definition": "label_values(project)",
-        "hide": 0,
         "includeAll": true,
         "multi": true,
         "name": "project",
@@ -863,43 +879,11 @@
         },
         "refresh": 1,
         "regex": "",
-        "skipUrlSync": false,
-        "sort": 0,
         "type": "query"
       },
       {
         "current": {
-          "selected": false,
-          "text": "dc1",
-          "value": "dc1"
-        },
-        "datasource": {
-          "datasource": "$datasource",
-          "uid": "$datasource"
-        },
-        "definition": "",
-        "hide": 0,
-        "includeAll": false,
-        "label": "DC",
-        "multi": false,
-        "name": "datacenter",
-        "options": [],
-        "query": "label_values(nomad_client_uptime, datacenter)",
-        "refresh": 1,
-        "regex": "",
-        "skipUrlSync": false,
-        "sort": 0,
-        "tagValuesQuery": "",
-        "tagsQuery": "",
-        "type": "query",
-        "useTags": false
-      },
-      {
-        "current": {
-          "selected": true,
-          "text": [
-            "All"
-          ],
+          "text": "All",
           "value": [
             "$__all"
           ]
@@ -908,61 +892,30 @@
           "datasource": "$datasource",
           "uid": "$datasource"
         },
-        "definition": "label_values(nomad_client_uptime{project=~\"$project\"},host)",
-        "hide": 0,
+        "definition": "label_values(nomad_client_uptime{project=~\"$project\"},instance)",
         "includeAll": true,
-        "label": "Host",
+        "label": "instance",
         "multi": true,
-        "name": "host",
+        "name": "instance",
         "options": [],
         "query": {
           "qryType": 1,
-          "query": "label_values(nomad_client_uptime{project=~\"$project\"},host)",
+          "query": "label_values(nomad_client_uptime{project=~\"$project\"},instance)",
           "refId": "PrometheusVariableQueryEditor-VariableQuery"
         },
         "refresh": 2,
         "regex": "",
-        "skipUrlSync": false,
-        "sort": 0,
-        "tagValuesQuery": "",
-        "tagsQuery": "",
-        "type": "query",
-        "useTags": false
+        "type": "query"
       }
     ]
   },
   "time": {
-    "from": "now-24h",
+    "from": "now-30m",
     "to": "now"
   },
-  "timepicker": {
-    "refresh_intervals": [
-      "5s",
-      "10s",
-      "30s",
-      "1m",
-      "5m",
-      "15m",
-      "30m",
-      "1h",
-      "2h",
-      "1d"
-    ],
-    "time_options": [
-      "5m",
-      "15m",
-      "1h",
-      "6h",
-      "12h",
-      "24h",
-      "2d",
-      "7d",
-      "30d"
-    ]
-  },
+  "timepicker": {},
   "timezone": "",
-  "title": "Nomad",
-  "uid": "ad9aff6e-226f-42ef-aa3f-8be3c80662ce",
-  "version": 1,
-  "weekStart": ""
-}
+  "title": "Nomad Copy2",
+  "uid": "hrvzvk7",
+  "version": 1
+}
\ No newline at end of file
diff --git a/ansible/playbooks/saas/roles/grafana/files/dashboards/nvidia/nvidia.json b/ansible/playbooks/saas/roles/grafana/files/dashboards/nvidia/nvidia.json
new file mode 100644
index 00000000..a5d821e7
--- /dev/null
+++ b/ansible/playbooks/saas/roles/grafana/files/dashboards/nvidia/nvidia.json
@@ -0,0 +1,2226 @@
+{
+  "__inputs": [
+    {
+      "name": "DS_PROMETHEUS",
+      "label": "Prometheus",
+      "description": "",
+      "type": "datasource",
+      "pluginId": "prometheus",
+      "pluginName": "Prometheus"
+    }
+  ],
+  "__elements": {},
+  "__requires": [
+    {
+      "type": "panel",
+      "id": "bargauge",
+      "name": "Bar gauge",
+      "version": ""
+    },
+    {
+      "type": "panel",
+      "id": "gauge",
+      "name": "Gauge",
+      "version": ""
+    },
+    {
+      "type": "grafana",
+      "id": "grafana",
+      "name": "Grafana",
+      "version": "11.2.0"
+    },
+    {
+      "type": "datasource",
+      "id": "prometheus",
+      "name": "Prometheus",
+      "version": "1.0.0"
+    },
+    {
+      "type": "panel",
+      "id": "stat",
+      "name": "Stat",
+      "version": ""
+    },
+    {
+      "type": "panel",
+      "id": "timeseries",
+      "name": "Time series",
+      "version": ""
+    }
+  ],
+  "annotations": {
+    "list": [
+      {
+        "builtIn": 1,
+        "datasource": {
+          "type": "datasource",
+          "uid": "grafana"
+        },
+        "enable": true,
+        "hide": true,
+        "iconColor": "rgba(0, 211, 255, 1)",
+        "name": "Annotations & Alerts",
+        "type": "dashboard"
+      }
+    ]
+  },
+  "description": "Nvidia GPU Metrics based on the prometheus metrics from github.com/utkuozdemir/nvidia_gpu_exporter",
+  "editable": true,
+  "fiscalYearStartMonth": 0,
+  "gnetId": 14574,
+  "graphTooltip": 0,
+  "id": null,
+  "links": [],
+  "panels": [
+    {
+      "datasource": {
+        "type": "prometheus",
+        "uid": "${DS_PROMETHEUS}"
+      },
+      "description": "The official product name of the GPU. This is an alphanumeric string. For all products.",
+      "fieldConfig": {
+        "defaults": {
+          "color": {
+            "mode": "thresholds"
+          },
+          "decimals": 2,
+          "mappings": [],
+          "thresholds": {
+            "mode": "absolute",
+            "steps": [
+              {
+                "color": "green",
+                "value": null
+              }
+            ]
+          },
+          "unit": "none"
+        },
+        "overrides": []
+      },
+      "gridPos": {
+        "h": 3,
+        "w": 4,
+        "x": 0,
+        "y": 0
+      },
+      "id": 23,
+      "options": {
+        "colorMode": "value",
+        "graphMode": "none",
+        "justifyMode": "auto",
+        "orientation": "auto",
+        "percentChangeColorMode": "standard",
+        "reduceOptions": {
+          "calcs": [
+            "last"
+          ],
+          "fields": "",
+          "values": false
+        },
+        "showPercentChange": false,
+        "text": {},
+        "textMode": "name",
+        "wideLayout": true
+      },
+      "pluginVersion": "11.2.0",
+      "targets": [
+        {
+          "datasource": {
+            "type": "prometheus",
+            "uid": "${DS_PROMETHEUS}"
+          },
+          "exemplar": true,
+          "expr": "nvidia_smi_gpu_info{uuid=\"$gpu\"}",
+          "instant": true,
+          "interval": "",
+          "legendFormat": "{{name}}",
+          "refId": "A"
+        }
+      ],
+      "title": "Name",
+      "type": "stat"
+    },
+    {
+      "datasource": {
+        "type": "prometheus",
+        "uid": "${DS_PROMETHEUS}"
+      },
+      "description": "The current performance state for the GPU. States range from P0 (maximum performance) to P12 (minimum performance).",
+      "fieldConfig": {
+        "defaults": {
+          "color": {
+            "mode": "thresholds"
+          },
+          "decimals": 0,
+          "mappings": [
+            {
+              "options": {
+                "": {
+                  "text": ""
+                }
+              },
+              "type": "value"
+            }
+          ],
+          "thresholds": {
+            "mode": "absolute",
+            "steps": [
+              {
+                "color": "green",
+                "value": null
+              }
+            ]
+          },
+          "unit": "prefix:P"
+        },
+        "overrides": []
+      },
+      "gridPos": {
+        "h": 3,
+        "w": 2,
+        "x": 4,
+        "y": 0
+      },
+      "id": 22,
+      "options": {
+        "colorMode": "value",
+        "graphMode": "none",
+        "justifyMode": "auto",
+        "orientation": "auto",
+        "percentChangeColorMode": "standard",
+        "reduceOptions": {
+          "calcs": [
+            "last"
+          ],
+          "fields": "",
+          "values": false
+        },
+        "showPercentChange": false,
+        "text": {},
+        "textMode": "value",
+        "wideLayout": true
+      },
+      "pluginVersion": "11.2.0",
+      "targets": [
+        {
+          "datasource": {
+            "type": "prometheus",
+            "uid": "${DS_PROMETHEUS}"
+          },
+          "exemplar": true,
+          "expr": "nvidia_smi_pstate{uuid=\"$gpu\"}",
+          "interval": "",
+          "legendFormat": "",
+          "refId": "A"
+        }
+      ],
+      "title": "P-State",
+      "type": "stat"
+    },
+    {
+      "datasource": {
+        "type": "prometheus",
+        "uid": "${DS_PROMETHEUS}"
+      },
+      "description": "Percent of time over the past sample period during which one or more kernels was executing on the GPU.\nThe sample period may be between 1 second and 1/6 second depending on the product.",
+      "fieldConfig": {
+        "defaults": {
+          "color": {
+            "mode": "thresholds"
+          },
+          "mappings": [],
+          "max": 1,
+          "min": 0,
+          "thresholds": {
+            "mode": "percentage",
+            "steps": [
+              {
+                "color": "green",
+                "value": null
+              },
+              {
+                "color": "#EAB839",
+                "value": 70
+              },
+              {
+                "color": "red",
+                "value": 90
+              }
+            ]
+          },
+          "unit": "percentunit"
+        },
+        "overrides": []
+      },
+      "gridPos": {
+        "h": 5,
+        "w": 3,
+        "x": 6,
+        "y": 0
+      },
+      "id": 6,
+      "options": {
+        "minVizHeight": 75,
+        "minVizWidth": 75,
+        "orientation": "auto",
+        "reduceOptions": {
+          "calcs": [
+            "last"
+          ],
+          "fields": "",
+          "values": false
+        },
+        "showThresholdLabels": false,
+        "showThresholdMarkers": true,
+        "sizing": "auto",
+        "text": {}
+      },
+      "pluginVersion": "11.2.0",
+      "targets": [
+        {
+          "datasource": {
+            "type": "prometheus",
+            "uid": "${DS_PROMETHEUS}"
+          },
+          "exemplar": true,
+          "expr": "nvidia_smi_utilization_gpu_ratio{uuid=\"$gpu\"}",
+          "interval": "",
+          "legendFormat": "{{uuid}}",
+          "refId": "A"
+        }
+      ],
+      "title": "GPU Utilization %",
+      "type": "gauge"
+    },
+    {
+      "datasource": {
+        "type": "prometheus",
+        "uid": "${DS_PROMETHEUS}"
+      },
+      "description": "The last measured power draw for the entire board, in watts. Only available if power management is supported. This reading is accurate to within +/- 5 watts / The software power limit in watts.",
+      "fieldConfig": {
+        "defaults": {
+          "color": {
+            "mode": "thresholds"
+          },
+          "mappings": [],
+          "max": 1,
+          "min": 0,
+          "thresholds": {
+            "mode": "percentage",
+            "steps": [
+              {
+                "color": "green",
+                "value": null
+              },
+              {
+                "color": "#EAB839",
+                "value": 70
+              },
+              {
+                "color": "red",
+                "value": 90
+              }
+            ]
+          },
+          "unit": "percentunit"
+        },
+        "overrides": []
+      },
+      "gridPos": {
+        "h": 5,
+        "w": 3,
+        "x": 9,
+        "y": 0
+      },
+      "id": 21,
+      "options": {
+        "minVizHeight": 75,
+        "minVizWidth": 75,
+        "orientation": "auto",
+        "reduceOptions": {
+          "calcs": [
+            "last"
+          ],
+          "fields": "",
+          "values": false
+        },
+        "showThresholdLabels": false,
+        "showThresholdMarkers": true,
+        "sizing": "auto",
+        "text": {}
+      },
+      "pluginVersion": "11.2.0",
+      "targets": [
+        {
+          "datasource": {
+            "type": "prometheus",
+            "uid": "${DS_PROMETHEUS}"
+          },
+          "exemplar": true,
+          "expr": "nvidia_smi_power_draw_watts{uuid=\"$gpu\"} / nvidia_smi_power_default_limit_watts{uuid=\"$gpu\"}",
+          "interval": "",
+          "legendFormat": "",
+          "refId": "A"
+        }
+      ],
+      "title": "Power Draw %",
+      "type": "gauge"
+    },
+    {
+      "datasource": {
+        "type": "prometheus",
+        "uid": "${DS_PROMETHEUS}"
+      },
+      "description": "The fan speed value is the percent of the product's maximum noise tolerance fan speed that the device's fan is currently intended to run at. This value may exceed 100% in certain cases. Note: The reported speed is the intended fan speed. If the fan is physically blocked and unable to spin, this output will not match the actual fan speed. Many parts do not report fan speeds because they rely on cooling via fans in the surrounding enclosure.\n",
+      "fieldConfig": {
+        "defaults": {
+          "color": {
+            "mode": "thresholds"
+          },
+          "mappings": [],
+          "max": 1,
+          "min": 0,
+          "thresholds": {
+            "mode": "percentage",
+            "steps": [
+              {
+                "color": "green",
+                "value": null
+              },
+              {
+                "color": "#EAB839",
+                "value": 70
+              },
+              {
+                "color": "red",
+                "value": 90
+              }
+            ]
+          },
+          "unit": "percentunit"
+        },
+        "overrides": []
+      },
+      "gridPos": {
+        "h": 5,
+        "w": 3,
+        "x": 12,
+        "y": 0
+      },
+      "id": 4,
+      "options": {
+        "minVizHeight": 75,
+        "minVizWidth": 75,
+        "orientation": "auto",
+        "reduceOptions": {
+          "calcs": [
+            "last"
+          ],
+          "fields": "",
+          "values": false
+        },
+        "showThresholdLabels": false,
+        "showThresholdMarkers": true,
+        "sizing": "auto",
+        "text": {}
+      },
+      "pluginVersion": "11.2.0",
+      "targets": [
+        {
+          "datasource": {
+            "type": "prometheus",
+            "uid": "${DS_PROMETHEUS}"
+          },
+          "exemplar": true,
+          "expr": "nvidia_smi_fan_speed_ratio{uuid=\"$gpu\"}",
+          "interval": "",
+          "legendFormat": "",
+          "refId": "A"
+        }
+      ],
+      "title": "Fan Speed %",
+      "type": "gauge"
+    },
+    {
+      "datasource": {
+        "type": "prometheus",
+        "uid": "${DS_PROMETHEUS}"
+      },
+      "description": "Core GPU temperature. in degrees C.",
+      "fieldConfig": {
+        "defaults": {
+          "color": {
+            "mode": "thresholds"
+          },
+          "mappings": [],
+          "max": 100,
+          "min": 0,
+          "thresholds": {
+            "mode": "absolute",
+            "steps": [
+              {
+                "color": "green",
+                "value": null
+              },
+              {
+                "color": "#EAB839",
+                "value": 70
+              },
+              {
+                "color": "red",
+                "value": 80
+              }
+            ]
+          },
+          "unit": "celsius"
+        },
+        "overrides": []
+      },
+      "gridPos": {
+        "h": 5,
+        "w": 3,
+        "x": 15,
+        "y": 0
+      },
+      "id": 16,
+      "options": {
+        "minVizHeight": 75,
+        "minVizWidth": 75,
+        "orientation": "auto",
+        "reduceOptions": {
+          "calcs": [
+            "last"
+          ],
+          "fields": "",
+          "values": false
+        },
+        "showThresholdLabels": false,
+        "showThresholdMarkers": true,
+        "sizing": "auto",
+        "text": {}
+      },
+      "pluginVersion": "11.2.0",
+      "targets": [
+        {
+          "datasource": {
+            "type": "prometheus",
+            "uid": "${DS_PROMETHEUS}"
+          },
+          "exemplar": true,
+          "expr": "nvidia_smi_temperature_gpu{uuid=\"$gpu\"}",
+          "interval": "",
+          "legendFormat": "{{uuid}}",
+          "refId": "A"
+        }
+      ],
+      "title": "Temperature",
+      "type": "gauge"
+    },
+    {
+      "datasource": {
+        "type": "prometheus",
+        "uid": "${DS_PROMETHEUS}"
+      },
+      "description": "Percent of time over the past sample period during which global (device) memory was being read or written.\nThe sample period may be between 1 second and 1/6 second depending on the product.",
+      "fieldConfig": {
+        "defaults": {
+          "color": {
+            "mode": "palette-classic"
+          },
+          "custom": {
+            "axisBorderShow": false,
+            "axisCenteredZero": false,
+            "axisColorMode": "text",
+            "axisLabel": "",
+            "axisPlacement": "auto",
+            "barAlignment": 0,
+            "barWidthFactor": 0.6,
+            "drawStyle": "line",
+            "fillOpacity": 10,
+            "gradientMode": "none",
+            "hideFrom": {
+              "legend": false,
+              "tooltip": false,
+              "viz": false
+            },
+            "insertNulls": false,
+            "lineInterpolation": "linear",
+            "lineWidth": 1,
+            "pointSize": 5,
+            "scaleDistribution": {
+              "type": "linear"
+            },
+            "showPoints": "never",
+            "spanNulls": false,
+            "stacking": {
+              "group": "A",
+              "mode": "none"
+            },
+            "thresholdsStyle": {
+              "mode": "line+area"
+            }
+          },
+          "mappings": [],
+          "max": 1,
+          "min": 0,
+          "thresholds": {
+            "mode": "absolute",
+            "steps": [
+              {
+                "color": "transparent",
+                "value": null
+              },
+              {
+                "color": "orange",
+                "value": 0.7
+              },
+              {
+                "color": "red",
+                "value": 0.9
+              }
+            ]
+          },
+          "unit": "percentunit"
+        },
+        "overrides": []
+      },
+      "gridPos": {
+        "h": 5,
+        "w": 6,
+        "x": 18,
+        "y": 0
+      },
+      "id": 11,
+      "options": {
+        "legend": {
+          "calcs": [],
+          "displayMode": "list",
+          "placement": "bottom",
+          "showLegend": false
+        },
+        "tooltip": {
+          "mode": "multi",
+          "sort": "none"
+        }
+      },
+      "pluginVersion": "11.1.0",
+      "targets": [
+        {
+          "datasource": {
+            "type": "prometheus",
+            "uid": "${DS_PROMETHEUS}"
+          },
+          "exemplar": true,
+          "expr": "nvidia_smi_utilization_memory_ratio{uuid=\"$gpu\"}",
+          "interval": "",
+          "legendFormat": "{{uuid}}",
+          "refId": "A"
+        }
+      ],
+      "title": "Memory Utilization %",
+      "type": "timeseries"
+    },
+    {
+      "datasource": {
+        "type": "prometheus",
+        "uid": "${DS_PROMETHEUS}"
+      },
+      "description": "The version of the installed NVIDIA display driver. This is an alphanumeric string.",
+      "fieldConfig": {
+        "defaults": {
+          "color": {
+            "mode": "thresholds"
+          },
+          "decimals": 2,
+          "mappings": [],
+          "thresholds": {
+            "mode": "absolute",
+            "steps": [
+              {
+                "color": "green",
+                "value": null
+              }
+            ]
+          },
+          "unit": "none"
+        },
+        "overrides": []
+      },
+      "gridPos": {
+        "h": 2,
+        "w": 3,
+        "x": 0,
+        "y": 3
+      },
+      "id": 14,
+      "options": {
+        "colorMode": "value",
+        "graphMode": "none",
+        "justifyMode": "auto",
+        "orientation": "auto",
+        "percentChangeColorMode": "standard",
+        "reduceOptions": {
+          "calcs": [
+            "last"
+          ],
+          "fields": "",
+          "values": false
+        },
+        "showPercentChange": false,
+        "text": {},
+        "textMode": "name",
+        "wideLayout": true
+      },
+      "pluginVersion": "11.2.0",
+      "targets": [
+        {
+          "datasource": {
+            "type": "prometheus",
+            "uid": "${DS_PROMETHEUS}"
+          },
+          "exemplar": true,
+          "expr": "nvidia_smi_gpu_info{uuid=\"$gpu\"}",
+          "instant": true,
+          "interval": "",
+          "legendFormat": "{{driver_version}}",
+          "refId": "A"
+        }
+      ],
+      "title": "Driver Version",
+      "type": "stat"
+    },
+    {
+      "datasource": {
+        "type": "prometheus",
+        "uid": "${DS_PROMETHEUS}"
+      },
+      "description": "The BIOS of the GPU board.",
+      "fieldConfig": {
+        "defaults": {
+          "color": {
+            "mode": "thresholds"
+          },
+          "decimals": 2,
+          "mappings": [],
+          "thresholds": {
+            "mode": "absolute",
+            "steps": [
+              {
+                "color": "green",
+                "value": null
+              }
+            ]
+          },
+          "unit": "none"
+        },
+        "overrides": []
+      },
+      "gridPos": {
+        "h": 2,
+        "w": 3,
+        "x": 3,
+        "y": 3
+      },
+      "id": 34,
+      "options": {
+        "colorMode": "value",
+        "graphMode": "none",
+        "justifyMode": "auto",
+        "orientation": "auto",
+        "percentChangeColorMode": "standard",
+        "reduceOptions": {
+          "calcs": [
+            "last"
+          ],
+          "fields": "",
+          "values": false
+        },
+        "showPercentChange": false,
+        "text": {},
+        "textMode": "name",
+        "wideLayout": true
+      },
+      "pluginVersion": "11.2.0",
+      "targets": [
+        {
+          "datasource": {
+            "type": "prometheus",
+            "uid": "${DS_PROMETHEUS}"
+          },
+          "exemplar": true,
+          "expr": "nvidia_smi_gpu_info{uuid=\"$gpu\"}",
+          "instant": true,
+          "interval": "",
+          "legendFormat": "{{vbios_version}}",
+          "refId": "A"
+        }
+      ],
+      "title": "Vbios Version",
+      "type": "stat"
+    },
+    {
+      "datasource": {
+        "type": "prometheus",
+        "uid": "${DS_PROMETHEUS}"
+      },
+      "description": "Information about factors that are reducing the frequency of clocks. If all throttle reasons are returned as \"Not Active\" it means that clocks are running as high as possible.",
+      "fieldConfig": {
+        "defaults": {
+          "color": {
+            "mode": "thresholds"
+          },
+          "decimals": 0,
+          "mappings": [
+            {
+              "options": {
+                "0": {
+                  "text": "Not Active"
+                },
+                "1": {
+                  "text": "Active"
+                }
+              },
+              "type": "value"
+            }
+          ],
+          "thresholds": {
+            "mode": "absolute",
+            "steps": [
+              {
+                "color": "green",
+                "value": null
+              }
+            ]
+          },
+          "unit": "none"
+        },
+        "overrides": []
+      },
+      "gridPos": {
+        "h": 5,
+        "w": 6,
+        "x": 0,
+        "y": 5
+      },
+      "id": 32,
+      "options": {
+        "displayMode": "gradient",
+        "maxVizHeight": 300,
+        "minVizHeight": 16,
+        "minVizWidth": 8,
+        "namePlacement": "auto",
+        "orientation": "horizontal",
+        "reduceOptions": {
+          "calcs": [
+            "last"
+          ],
+          "fields": "",
+          "values": false
+        },
+        "showUnfilled": true,
+        "sizing": "auto",
+        "text": {},
+        "valueMode": "color"
+      },
+      "pluginVersion": "11.2.0",
+      "targets": [
+        {
+          "datasource": {
+            "type": "prometheus",
+            "uid": "${DS_PROMETHEUS}"
+          },
+          "exemplar": true,
+          "expr": "nvidia_smi_clocks_event_reasons_gpu_idle{uuid=\"$gpu\"} or nvidia_smi_clocks_throttle_reasons_gpu_idle{uuid=\"$gpu\"}",
+          "instant": false,
+          "interval": "",
+          "legendFormat": "Idle",
+          "refId": "A"
+        },
+        {
+          "datasource": {
+            "type": "prometheus",
+            "uid": "${DS_PROMETHEUS}"
+          },
+          "exemplar": true,
+          "expr": "nvidia_smi_clocks_event_reasons_hw_thermal_slowdown{uuid=\"$gpu\"} or nvidia_smi_clocks_throttle_reasons_hw_thermal_slowdown{uuid=\"$gpu\"}",
+          "hide": false,
+          "interval": "",
+          "legendFormat": "HW Thermal Slowdown",
+          "refId": "B"
+        },
+        {
+          "datasource": {
+            "type": "prometheus",
+            "uid": "${DS_PROMETHEUS}"
+          },
+          "exemplar": true,
+          "expr": "nvidia_smi_clocks_event_reasons_sw_power_cap{uuid=\"$gpu\"} or nvidia_smi_clocks_throttle_reasons_sw_power_cap{uuid=\"$gpu\"}",
+          "hide": false,
+          "interval": "",
+          "legendFormat": "SW Power Cap",
+          "refId": "C"
+        },
+        {
+          "datasource": {
+            "type": "prometheus",
+            "uid": "${DS_PROMETHEUS}"
+          },
+          "exemplar": true,
+          "expr": "nvidia_smi_clocks_event_reasons_applications_clocks_setting{uuid=\"$gpu\"} or nvidia_smi_clocks_throttle_reasons_applications_clocks_setting{uuid=\"$gpu\"}",
+          "hide": false,
+          "interval": "",
+          "legendFormat": "App Clocks Setting",
+          "refId": "D"
+        },
+        {
+          "datasource": {
+            "type": "prometheus",
+            "uid": "${DS_PROMETHEUS}"
+          },
+          "exemplar": true,
+          "expr": "nvidia_smi_clocks_event_reasons_hw_power_brake_slowdown{uuid=\"$gpu\"} or nvidia_smi_clocks_throttle_reasons_hw_power_brake_slowdown{uuid=\"$gpu\"}",
+          "hide": false,
+          "interval": "",
+          "legendFormat": "HW Power Brake",
+          "refId": "E"
+        },
+        {
+          "datasource": {
+            "type": "prometheus",
+            "uid": "${DS_PROMETHEUS}"
+          },
+          "exemplar": true,
+          "expr": "nvidia_smi_clocks_event_reasons_sw_thermal_slowdown{uuid=\"$gpu\"} or nvidia_smi_clocks_throttle_reasons_sw_thermal_slowdown{uuid=\"$gpu\"}",
+          "hide": false,
+          "interval": "",
+          "legendFormat": "SW Thermal Slowdown",
+          "refId": "F"
+        },
+        {
+          "datasource": {
+            "type": "prometheus",
+            "uid": "${DS_PROMETHEUS}"
+          },
+          "exemplar": true,
+          "expr": "nvidia_smi_clocks_event_reasons_sync_boost{uuid=\"$gpu\"} or nvidia_smi_clocks_throttle_reasons_sync_boost{uuid=\"$gpu\"}",
+          "hide": false,
+          "interval": "",
+          "legendFormat": "Sync Boost",
+          "refId": "G"
+        }
+      ],
+      "title": "Throttle Reasons",
+      "type": "bargauge"
+    },
+    {
+      "datasource": {
+        "type": "prometheus",
+        "uid": "${DS_PROMETHEUS}"
+      },
+      "description": "Current frequency of graphics (shader) clock\n/\nMaximum frequency of graphics (shader) clock.\n",
+      "fieldConfig": {
+        "defaults": {
+          "color": {
+            "mode": "thresholds"
+          },
+          "mappings": [],
+          "max": 1,
+          "min": 0,
+          "thresholds": {
+            "mode": "percentage",
+            "steps": [
+              {
+                "color": "green",
+                "value": null
+              },
+              {
+                "color": "#EAB839",
+                "value": 70
+              },
+              {
+                "color": "red",
+                "value": 90
+              }
+            ]
+          },
+          "unit": "percentunit"
+        },
+        "overrides": []
+      },
+      "gridPos": {
+        "h": 5,
+        "w": 3,
+        "x": 6,
+        "y": 5
+      },
+      "id": 20,
+      "options": {
+        "minVizHeight": 75,
+        "minVizWidth": 75,
+        "orientation": "auto",
+        "reduceOptions": {
+          "calcs": [
+            "last"
+          ],
+          "fields": "",
+          "values": false
+        },
+        "showThresholdLabels": false,
+        "showThresholdMarkers": true,
+        "sizing": "auto",
+        "text": {}
+      },
+      "pluginVersion": "11.2.0",
+      "targets": [
+        {
+          "datasource": {
+            "type": "prometheus",
+            "uid": "${DS_PROMETHEUS}"
+          },
+          "exemplar": true,
+          "expr": "nvidia_smi_clocks_current_graphics_clock_hz{uuid=\"$gpu\"} / nvidia_smi_clocks_max_graphics_clock_hz{uuid=\"$gpu\"}",
+          "interval": "",
+          "legendFormat": "",
+          "refId": "A"
+        }
+      ],
+      "title": "GPU Clock Speed %",
+      "type": "gauge"
+    },
+    {
+      "datasource": {
+        "type": "prometheus",
+        "uid": "${DS_PROMETHEUS}"
+      },
+      "description": "Current frequency of memory clock / Maximum frequency of memory clock",
+      "fieldConfig": {
+        "defaults": {
+          "color": {
+            "mode": "thresholds"
+          },
+          "mappings": [],
+          "max": 1,
+          "min": 0,
+          "thresholds": {
+            "mode": "percentage",
+            "steps": [
+              {
+                "color": "green",
+                "value": null
+              },
+              {
+                "color": "#EAB839",
+                "value": 70
+              },
+              {
+                "color": "red",
+                "value": 90
+              }
+            ]
+          },
+          "unit": "percentunit"
+        },
+        "overrides": []
+      },
+      "gridPos": {
+        "h": 5,
+        "w": 3,
+        "x": 9,
+        "y": 5
+      },
+      "id": 33,
+      "options": {
+        "minVizHeight": 75,
+        "minVizWidth": 75,
+        "orientation": "auto",
+        "reduceOptions": {
+          "calcs": [
+            "last"
+          ],
+          "fields": "",
+          "values": false
+        },
+        "showThresholdLabels": false,
+        "showThresholdMarkers": true,
+        "sizing": "auto",
+        "text": {}
+      },
+      "pluginVersion": "11.2.0",
+      "targets": [
+        {
+          "datasource": {
+            "type": "prometheus",
+            "uid": "${DS_PROMETHEUS}"
+          },
+          "exemplar": true,
+          "expr": "nvidia_smi_clocks_current_memory_clock_hz{uuid=\"$gpu\"} / nvidia_smi_clocks_max_memory_clock_hz{uuid=\"$gpu\"}",
+          "interval": "",
+          "legendFormat": "",
+          "refId": "A"
+        }
+      ],
+      "title": "Memory Clock Speed %",
+      "type": "gauge"
+    },
+    {
+      "datasource": {
+        "type": "prometheus",
+        "uid": "${DS_PROMETHEUS}"
+      },
+      "description": "Total memory allocated by active contexts / Total installed GPU memory.",
+      "fieldConfig": {
+        "defaults": {
+          "color": {
+            "mode": "thresholds"
+          },
+          "mappings": [],
+          "max": 1,
+          "min": 0,
+          "thresholds": {
+            "mode": "percentage",
+            "steps": [
+              {
+                "color": "green",
+                "value": null
+              },
+              {
+                "color": "#EAB839",
+                "value": 70
+              },
+              {
+                "color": "red",
+                "value": 90
+              }
+            ]
+          },
+          "unit": "percentunit"
+        },
+        "overrides": []
+      },
+      "gridPos": {
+        "h": 5,
+        "w": 3,
+        "x": 12,
+        "y": 5
+      },
+      "id": 25,
+      "options": {
+        "minVizHeight": 75,
+        "minVizWidth": 75,
+        "orientation": "auto",
+        "reduceOptions": {
+          "calcs": [
+            "last"
+          ],
+          "fields": "",
+          "values": false
+        },
+        "showThresholdLabels": false,
+        "showThresholdMarkers": true,
+        "sizing": "auto",
+        "text": {}
+      },
+      "pluginVersion": "11.2.0",
+      "targets": [
+        {
+          "datasource": {
+            "type": "prometheus",
+            "uid": "${DS_PROMETHEUS}"
+          },
+          "exemplar": true,
+          "expr": "nvidia_smi_memory_used_bytes{uuid=\"$gpu\"} / nvidia_smi_memory_total_bytes{uuid=\"$gpu\"}",
+          "interval": "",
+          "legendFormat": "",
+          "refId": "A"
+        }
+      ],
+      "title": "Memory Allocation %",
+      "type": "gauge"
+    },
+    {
+      "datasource": {
+        "type": "prometheus",
+        "uid": "${DS_PROMETHEUS}"
+      },
+      "description": "Percent of time over the past sample period during which global (device) memory was being read or written.\nThe sample period may be between 1 second and 1/6 second depending on the product.",
+      "fieldConfig": {
+        "defaults": {
+          "color": {
+            "mode": "thresholds"
+          },
+          "mappings": [],
+          "max": 1,
+          "min": 0,
+          "thresholds": {
+            "mode": "percentage",
+            "steps": [
+              {
+                "color": "green",
+                "value": null
+              },
+              {
+                "color": "#EAB839",
+                "value": 70
+              },
+              {
+                "color": "red",
+                "value": 90
+              }
+            ]
+          },
+          "unit": "percentunit"
+        },
+        "overrides": []
+      },
+      "gridPos": {
+        "h": 5,
+        "w": 3,
+        "x": 15,
+        "y": 5
+      },
+      "id": 7,
+      "options": {
+        "minVizHeight": 75,
+        "minVizWidth": 75,
+        "orientation": "auto",
+        "reduceOptions": {
+          "calcs": [
+            "last"
+          ],
+          "fields": "",
+          "values": false
+        },
+        "showThresholdLabels": false,
+        "showThresholdMarkers": true,
+        "sizing": "auto",
+        "text": {}
+      },
+      "pluginVersion": "11.2.0",
+      "targets": [
+        {
+          "datasource": {
+            "type": "prometheus",
+            "uid": "${DS_PROMETHEUS}"
+          },
+          "exemplar": true,
+          "expr": "nvidia_smi_utilization_memory_ratio{uuid=\"$gpu\"}",
+          "interval": "",
+          "legendFormat": "",
+          "refId": "A"
+        }
+      ],
+      "title": "Memory Utilization %",
+      "type": "gauge"
+    },
+    {
+      "datasource": {
+        "type": "prometheus",
+        "uid": "${DS_PROMETHEUS}"
+      },
+      "description": "Percent of time over the past sample period during which one or more kernels was executing on the GPU.\nThe sample period may be between 1 second and 1/6 second depending on the product.",
+      "fieldConfig": {
+        "defaults": {
+          "color": {
+            "mode": "palette-classic"
+          },
+          "custom": {
+            "axisBorderShow": false,
+            "axisCenteredZero": false,
+            "axisColorMode": "text",
+            "axisLabel": "",
+            "axisPlacement": "auto",
+            "barAlignment": 0,
+            "barWidthFactor": 0.6,
+            "drawStyle": "line",
+            "fillOpacity": 10,
+            "gradientMode": "none",
+            "hideFrom": {
+              "legend": false,
+              "tooltip": false,
+              "viz": false
+            },
+            "insertNulls": false,
+            "lineInterpolation": "linear",
+            "lineWidth": 1,
+            "pointSize": 5,
+            "scaleDistribution": {
+              "type": "linear"
+            },
+            "showPoints": "never",
+            "spanNulls": false,
+            "stacking": {
+              "group": "A",
+              "mode": "none"
+            },
+            "thresholdsStyle": {
+              "mode": "line+area"
+            }
+          },
+          "mappings": [],
+          "max": 1,
+          "min": 0,
+          "thresholds": {
+            "mode": "absolute",
+            "steps": [
+              {
+                "color": "transparent",
+                "value": null
+              },
+              {
+                "color": "orange",
+                "value": 0.7
+              },
+              {
+                "color": "red",
+                "value": 0.9
+              }
+            ]
+          },
+          "unit": "percentunit"
+        },
+        "overrides": []
+      },
+      "gridPos": {
+        "h": 5,
+        "w": 6,
+        "x": 18,
+        "y": 5
+      },
+      "id": 10,
+      "options": {
+        "legend": {
+          "calcs": [],
+          "displayMode": "list",
+          "placement": "bottom",
+          "showLegend": false
+        },
+        "tooltip": {
+          "mode": "multi",
+          "sort": "none"
+        }
+      },
+      "pluginVersion": "11.1.0",
+      "targets": [
+        {
+          "datasource": {
+            "type": "prometheus",
+            "uid": "${DS_PROMETHEUS}"
+          },
+          "exemplar": true,
+          "expr": "nvidia_smi_utilization_gpu_ratio{uuid=\"$gpu\"}",
+          "interval": "",
+          "legendFormat": "",
+          "refId": "A"
+        }
+      ],
+      "title": "GPU Utilization %",
+      "type": "timeseries"
+    },
+    {
+      "datasource": {
+        "type": "prometheus",
+        "uid": "${DS_PROMETHEUS}"
+      },
+      "description": "Total memory allocated by active contexts.",
+      "fieldConfig": {
+        "defaults": {
+          "color": {
+            "mode": "palette-classic"
+          },
+          "custom": {
+            "axisBorderShow": false,
+            "axisCenteredZero": false,
+            "axisColorMode": "text",
+            "axisLabel": "",
+            "axisPlacement": "auto",
+            "barAlignment": 0,
+            "barWidthFactor": 0.6,
+            "drawStyle": "line",
+            "fillOpacity": 10,
+            "gradientMode": "none",
+            "hideFrom": {
+              "legend": false,
+              "tooltip": false,
+              "viz": false
+            },
+            "insertNulls": false,
+            "lineInterpolation": "linear",
+            "lineWidth": 1,
+            "pointSize": 5,
+            "scaleDistribution": {
+              "type": "linear"
+            },
+            "showPoints": "never",
+            "spanNulls": false,
+            "stacking": {
+              "group": "A",
+              "mode": "none"
+            },
+            "thresholdsStyle": {
+              "mode": "off"
+            }
+          },
+          "mappings": [],
+          "min": 0,
+          "thresholds": {
+            "mode": "absolute",
+            "steps": [
+              {
+                "color": "green",
+                "value": null
+              },
+              {
+                "color": "red",
+                "value": 80
+              }
+            ]
+          },
+          "unit": "decbytes"
+        },
+        "overrides": []
+      },
+      "gridPos": {
+        "h": 5,
+        "w": 6,
+        "x": 0,
+        "y": 10
+      },
+      "id": 17,
+      "options": {
+        "legend": {
+          "calcs": [],
+          "displayMode": "list",
+          "placement": "bottom",
+          "showLegend": false
+        },
+        "tooltip": {
+          "mode": "multi",
+          "sort": "none"
+        }
+      },
+      "pluginVersion": "11.1.0",
+      "targets": [
+        {
+          "datasource": {
+            "type": "prometheus",
+            "uid": "${DS_PROMETHEUS}"
+          },
+          "exemplar": true,
+          "expr": "nvidia_smi_memory_used_bytes{uuid=\"$gpu\"}",
+          "interval": "",
+          "legendFormat": "{{uuid}}",
+          "refId": "A"
+        }
+      ],
+      "title": "Memory Allocation",
+      "type": "timeseries"
+    },
+    {
+      "datasource": {
+        "type": "prometheus",
+        "uid": "${DS_PROMETHEUS}"
+      },
+      "description": "Core GPU temperature. in degrees C.",
+      "fieldConfig": {
+        "defaults": {
+          "color": {
+            "mode": "palette-classic"
+          },
+          "custom": {
+            "axisBorderShow": false,
+            "axisCenteredZero": false,
+            "axisColorMode": "text",
+            "axisLabel": "",
+            "axisPlacement": "auto",
+            "barAlignment": 0,
+            "barWidthFactor": 0.6,
+            "drawStyle": "line",
+            "fillOpacity": 10,
+            "gradientMode": "none",
+            "hideFrom": {
+              "legend": false,
+              "tooltip": false,
+              "viz": false
+            },
+            "insertNulls": false,
+            "lineInterpolation": "linear",
+            "lineWidth": 1,
+            "pointSize": 5,
+            "scaleDistribution": {
+              "type": "linear"
+            },
+            "showPoints": "never",
+            "spanNulls": false,
+            "stacking": {
+              "group": "A",
+              "mode": "none"
+            },
+            "thresholdsStyle": {
+              "mode": "line+area"
+            }
+          },
+          "mappings": [],
+          "max": 100,
+          "min": 0,
+          "thresholds": {
+            "mode": "absolute",
+            "steps": [
+              {
+                "color": "transparent",
+                "value": null
+              },
+              {
+                "color": "orange",
+                "value": 70
+              },
+              {
+                "color": "red",
+                "value": 80
+              }
+            ]
+          },
+          "unit": "celsius"
+        },
+        "overrides": []
+      },
+      "gridPos": {
+        "h": 5,
+        "w": 6,
+        "x": 6,
+        "y": 10
+      },
+      "id": 15,
+      "options": {
+        "legend": {
+          "calcs": [],
+          "displayMode": "list",
+          "placement": "bottom",
+          "showLegend": false
+        },
+        "tooltip": {
+          "mode": "multi",
+          "sort": "none"
+        }
+      },
+      "pluginVersion": "11.1.0",
+      "targets": [
+        {
+          "datasource": {
+            "type": "prometheus",
+            "uid": "${DS_PROMETHEUS}"
+          },
+          "exemplar": true,
+          "expr": "nvidia_smi_temperature_gpu{uuid=\"$gpu\"}",
+          "interval": "",
+          "legendFormat": "{{uuid}}",
+          "refId": "A"
+        }
+      ],
+      "title": "Temperature",
+      "type": "timeseries"
+    },
+    {
+      "datasource": {
+        "type": "prometheus",
+        "uid": "${DS_PROMETHEUS}"
+      },
+      "description": "The last measured power draw for the entire board, in watts. Only available if power management is supported. This reading is accurate to within +/- 5 watts",
+      "fieldConfig": {
+        "defaults": {
+          "color": {
+            "mode": "palette-classic"
+          },
+          "custom": {
+            "axisBorderShow": false,
+            "axisCenteredZero": false,
+            "axisColorMode": "text",
+            "axisLabel": "",
+            "axisPlacement": "auto",
+            "barAlignment": 0,
+            "barWidthFactor": 0.6,
+            "drawStyle": "line",
+            "fillOpacity": 10,
+            "gradientMode": "none",
+            "hideFrom": {
+              "legend": false,
+              "tooltip": false,
+              "viz": false
+            },
+            "insertNulls": false,
+            "lineInterpolation": "linear",
+            "lineWidth": 1,
+            "pointSize": 5,
+            "scaleDistribution": {
+              "type": "linear"
+            },
+            "showPoints": "never",
+            "spanNulls": false,
+            "stacking": {
+              "group": "A",
+              "mode": "none"
+            },
+            "thresholdsStyle": {
+              "mode": "off"
+            }
+          },
+          "mappings": [],
+          "min": 0,
+          "thresholds": {
+            "mode": "absolute",
+            "steps": [
+              {
+                "color": "green",
+                "value": null
+              },
+              {
+                "color": "red",
+                "value": 80
+              }
+            ]
+          },
+          "unit": "watt"
+        },
+        "overrides": []
+      },
+      "gridPos": {
+        "h": 5,
+        "w": 6,
+        "x": 12,
+        "y": 10
+      },
+      "id": 8,
+      "options": {
+        "legend": {
+          "calcs": [],
+          "displayMode": "list",
+          "placement": "bottom",
+          "showLegend": false
+        },
+        "tooltip": {
+          "mode": "multi",
+          "sort": "none"
+        }
+      },
+      "pluginVersion": "11.1.0",
+      "targets": [
+        {
+          "datasource": {
+            "type": "prometheus",
+            "uid": "${DS_PROMETHEUS}"
+          },
+          "exemplar": true,
+          "expr": "nvidia_smi_power_draw_watts{uuid=\"$gpu\"}",
+          "interval": "",
+          "legendFormat": "{{uuid}}",
+          "refId": "A"
+        }
+      ],
+      "title": "Power Draw",
+      "type": "timeseries"
+    },
+    {
+      "datasource": {
+        "type": "prometheus",
+        "uid": "${DS_PROMETHEUS}"
+      },
+      "description": "The fan speed value is the percent of the product's maximum noise tolerance fan speed that the device's fan is currently intended to run at. This value may exceed 100% in certain cases. Note: The reported speed is the intended fan speed. If the fan is physically blocked and unable to spin, this output will not match the actual fan speed. Many parts do not report fan speeds because they rely on cooling via fans in the surrounding enclosure.\n",
+      "fieldConfig": {
+        "defaults": {
+          "color": {
+            "mode": "palette-classic"
+          },
+          "custom": {
+            "axisBorderShow": false,
+            "axisCenteredZero": false,
+            "axisColorMode": "text",
+            "axisLabel": "",
+            "axisPlacement": "auto",
+            "barAlignment": 0,
+            "barWidthFactor": 0.6,
+            "drawStyle": "line",
+            "fillOpacity": 10,
+            "gradientMode": "none",
+            "hideFrom": {
+              "legend": false,
+              "tooltip": false,
+              "viz": false
+            },
+            "insertNulls": false,
+            "lineInterpolation": "linear",
+            "lineWidth": 1,
+            "pointSize": 5,
+            "scaleDistribution": {
+              "type": "linear"
+            },
+            "showPoints": "never",
+            "spanNulls": false,
+            "stacking": {
+              "group": "A",
+              "mode": "none"
+            },
+            "thresholdsStyle": {
+              "mode": "line+area"
+            }
+          },
+          "mappings": [],
+          "max": 1,
+          "min": 0,
+          "thresholds": {
+            "mode": "absolute",
+            "steps": [
+              {
+                "color": "transparent",
+                "value": null
+              },
+              {
+                "color": "orange",
+                "value": 0.7
+              },
+              {
+                "color": "red",
+                "value": 0.9
+              }
+            ]
+          },
+          "unit": "percentunit"
+        },
+        "overrides": []
+      },
+      "gridPos": {
+        "h": 5,
+        "w": 6,
+        "x": 18,
+        "y": 10
+      },
+      "id": 9,
+      "options": {
+        "legend": {
+          "calcs": [],
+          "displayMode": "list",
+          "placement": "bottom",
+          "showLegend": false
+        },
+        "tooltip": {
+          "mode": "multi",
+          "sort": "none"
+        }
+      },
+      "pluginVersion": "11.1.0",
+      "targets": [
+        {
+          "datasource": {
+            "type": "prometheus",
+            "uid": "${DS_PROMETHEUS}"
+          },
+          "exemplar": true,
+          "expr": "nvidia_smi_fan_speed_ratio{uuid=\"$gpu\"}",
+          "interval": "",
+          "legendFormat": "{{uuid}}",
+          "refId": "A"
+        }
+      ],
+      "title": "Fan Speed %",
+      "type": "timeseries"
+    },
+    {
+      "datasource": {
+        "type": "prometheus",
+        "uid": "${DS_PROMETHEUS}"
+      },
+      "description": "Current frequency of graphics (shader) clock.",
+      "fieldConfig": {
+        "defaults": {
+          "color": {
+            "mode": "palette-classic"
+          },
+          "custom": {
+            "axisBorderShow": false,
+            "axisCenteredZero": false,
+            "axisColorMode": "text",
+            "axisLabel": "",
+            "axisPlacement": "auto",
+            "barAlignment": 0,
+            "barWidthFactor": 0.6,
+            "drawStyle": "line",
+            "fillOpacity": 10,
+            "gradientMode": "none",
+            "hideFrom": {
+              "legend": false,
+              "tooltip": false,
+              "viz": false
+            },
+            "insertNulls": false,
+            "lineInterpolation": "linear",
+            "lineWidth": 1,
+            "pointSize": 5,
+            "scaleDistribution": {
+              "type": "linear"
+            },
+            "showPoints": "never",
+            "spanNulls": false,
+            "stacking": {
+              "group": "A",
+              "mode": "none"
+            },
+            "thresholdsStyle": {
+              "mode": "off"
+            }
+          },
+          "mappings": [],
+          "min": 0,
+          "thresholds": {
+            "mode": "absolute",
+            "steps": [
+              {
+                "color": "green",
+                "value": null
+              },
+              {
+                "color": "red",
+                "value": 80
+              }
+            ]
+          },
+          "unit": "hertz"
+        },
+        "overrides": []
+      },
+      "gridPos": {
+        "h": 5,
+        "w": 6,
+        "x": 0,
+        "y": 15
+      },
+      "id": 12,
+      "options": {
+        "legend": {
+          "calcs": [],
+          "displayMode": "list",
+          "placement": "bottom",
+          "showLegend": false
+        },
+        "tooltip": {
+          "mode": "multi",
+          "sort": "none"
+        }
+      },
+      "pluginVersion": "11.1.0",
+      "targets": [
+        {
+          "datasource": {
+            "type": "prometheus",
+            "uid": "${DS_PROMETHEUS}"
+          },
+          "exemplar": true,
+          "expr": "nvidia_smi_clocks_current_graphics_clock_hz{uuid=\"$gpu\"}",
+          "format": "time_series",
+          "interval": "",
+          "legendFormat": "{{uuid}}",
+          "refId": "A"
+        }
+      ],
+      "title": "Graphics Clock Speed",
+      "type": "timeseries"
+    },
+    {
+      "datasource": {
+        "type": "prometheus",
+        "uid": "${DS_PROMETHEUS}"
+      },
+      "description": "Current frequency of video encoder/decoder clock.",
+      "fieldConfig": {
+        "defaults": {
+          "color": {
+            "mode": "palette-classic"
+          },
+          "custom": {
+            "axisBorderShow": false,
+            "axisCenteredZero": false,
+            "axisColorMode": "text",
+            "axisLabel": "",
+            "axisPlacement": "auto",
+            "barAlignment": 0,
+            "barWidthFactor": 0.6,
+            "drawStyle": "line",
+            "fillOpacity": 10,
+            "gradientMode": "none",
+            "hideFrom": {
+              "legend": false,
+              "tooltip": false,
+              "viz": false
+            },
+            "insertNulls": false,
+            "lineInterpolation": "linear",
+            "lineWidth": 1,
+            "pointSize": 5,
+            "scaleDistribution": {
+              "type": "linear"
+            },
+            "showPoints": "never",
+            "spanNulls": false,
+            "stacking": {
+              "group": "A",
+              "mode": "none"
+            },
+            "thresholdsStyle": {
+              "mode": "off"
+            }
+          },
+          "mappings": [],
+          "min": 0,
+          "thresholds": {
+            "mode": "absolute",
+            "steps": [
+              {
+                "color": "green",
+                "value": null
+              },
+              {
+                "color": "red",
+                "value": 80
+              }
+            ]
+          },
+          "unit": "hertz"
+        },
+        "overrides": []
+      },
+      "gridPos": {
+        "h": 5,
+        "w": 6,
+        "x": 6,
+        "y": 15
+      },
+      "id": 19,
+      "options": {
+        "legend": {
+          "calcs": [],
+          "displayMode": "list",
+          "placement": "bottom",
+          "showLegend": false
+        },
+        "tooltip": {
+          "mode": "multi",
+          "sort": "none"
+        }
+      },
+      "pluginVersion": "11.1.0",
+      "targets": [
+        {
+          "datasource": {
+            "type": "prometheus",
+            "uid": "${DS_PROMETHEUS}"
+          },
+          "exemplar": true,
+          "expr": "nvidia_smi_clocks_current_video_clock_hz{uuid=\"$gpu\"}",
+          "format": "time_series",
+          "interval": "",
+          "legendFormat": "{{uuid}}",
+          "refId": "A"
+        }
+      ],
+      "title": "Video Clock Speed",
+      "type": "timeseries"
+    },
+    {
+      "datasource": {
+        "type": "prometheus",
+        "uid": "${DS_PROMETHEUS}"
+      },
+      "description": "Current frequency of SM (Streaming Multiprocessor) clock.",
+      "fieldConfig": {
+        "defaults": {
+          "color": {
+            "mode": "palette-classic"
+          },
+          "custom": {
+            "axisBorderShow": false,
+            "axisCenteredZero": false,
+            "axisColorMode": "text",
+            "axisLabel": "",
+            "axisPlacement": "auto",
+            "barAlignment": 0,
+            "barWidthFactor": 0.6,
+            "drawStyle": "line",
+            "fillOpacity": 10,
+            "gradientMode": "none",
+            "hideFrom": {
+              "legend": false,
+              "tooltip": false,
+              "viz": false
+            },
+            "insertNulls": false,
+            "lineInterpolation": "linear",
+            "lineWidth": 1,
+            "pointSize": 5,
+            "scaleDistribution": {
+              "type": "linear"
+            },
+            "showPoints": "never",
+            "spanNulls": false,
+            "stacking": {
+              "group": "A",
+              "mode": "none"
+            },
+            "thresholdsStyle": {
+              "mode": "off"
+            }
+          },
+          "mappings": [],
+          "min": 0,
+          "thresholds": {
+            "mode": "absolute",
+            "steps": [
+              {
+                "color": "green",
+                "value": null
+              },
+              {
+                "color": "red",
+                "value": 80
+              }
+            ]
+          },
+          "unit": "hertz"
+        },
+        "overrides": []
+      },
+      "gridPos": {
+        "h": 5,
+        "w": 6,
+        "x": 12,
+        "y": 15
+      },
+      "id": 24,
+      "options": {
+        "legend": {
+          "calcs": [],
+          "displayMode": "list",
+          "placement": "bottom",
+          "showLegend": false
+        },
+        "tooltip": {
+          "mode": "multi",
+          "sort": "none"
+        }
+      },
+      "pluginVersion": "11.1.0",
+      "targets": [
+        {
+          "datasource": {
+            "type": "prometheus",
+            "uid": "${DS_PROMETHEUS}"
+          },
+          "exemplar": true,
+          "expr": "nvidia_smi_clocks_current_sm_clock_hz{uuid=\"$gpu\"}",
+          "format": "time_series",
+          "interval": "",
+          "legendFormat": "{{uuid}}",
+          "refId": "A"
+        }
+      ],
+      "title": "SM Clock Speed",
+      "type": "timeseries"
+    },
+    {
+      "datasource": {
+        "type": "prometheus",
+        "uid": "${DS_PROMETHEUS}"
+      },
+      "description": "Current frequency of memory clock.",
+      "fieldConfig": {
+        "defaults": {
+          "color": {
+            "mode": "palette-classic"
+          },
+          "custom": {
+            "axisBorderShow": false,
+            "axisCenteredZero": false,
+            "axisColorMode": "text",
+            "axisLabel": "",
+            "axisPlacement": "auto",
+            "barAlignment": 0,
+            "barWidthFactor": 0.6,
+            "drawStyle": "line",
+            "fillOpacity": 10,
+            "gradientMode": "none",
+            "hideFrom": {
+              "legend": false,
+              "tooltip": false,
+              "viz": false
+            },
+            "insertNulls": false,
+            "lineInterpolation": "linear",
+            "lineWidth": 1,
+            "pointSize": 5,
+            "scaleDistribution": {
+              "type": "linear"
+            },
+            "showPoints": "never",
+            "spanNulls": false,
+            "stacking": {
+              "group": "A",
+              "mode": "none"
+            },
+            "thresholdsStyle": {
+              "mode": "off"
+            }
+          },
+          "mappings": [],
+          "min": 0,
+          "thresholds": {
+            "mode": "absolute",
+            "steps": [
+              {
+                "color": "green",
+                "value": null
+              },
+              {
+                "color": "red",
+                "value": 80
+              }
+            ]
+          },
+          "unit": "hertz"
+        },
+        "overrides": []
+      },
+      "gridPos": {
+        "h": 5,
+        "w": 6,
+        "x": 18,
+        "y": 15
+      },
+      "id": 18,
+      "options": {
+        "legend": {
+          "calcs": [],
+          "displayMode": "list",
+          "placement": "bottom",
+          "showLegend": false
+        },
+        "tooltip": {
+          "mode": "multi",
+          "sort": "none"
+        }
+      },
+      "pluginVersion": "11.1.0",
+      "targets": [
+        {
+          "datasource": {
+            "type": "prometheus",
+            "uid": "${DS_PROMETHEUS}"
+          },
+          "exemplar": true,
+          "expr": "nvidia_smi_clocks_current_memory_clock_hz{uuid=\"$gpu\"}",
+          "format": "time_series",
+          "interval": "",
+          "legendFormat": "{{uuid}}",
+          "refId": "A"
+        }
+      ],
+      "title": "Memory Clock Speed",
+      "type": "timeseries"
+    }
+  ],
+  "refresh": "10s",
+  "schemaVersion": 39,
+  "tags": [
+    "nvidia",
+    "nvidia-smi",
+    "nvidia_gpu_exporter",
+    "prometheus"
+  ],
+  "templating": {
+    "list": [
+      {
+        "current": {},
+        "datasource": {
+          "type": "prometheus",
+          "uid": "${DS_PROMETHEUS}"
+        },
+        "definition": "label_values(nvidia_smi_index, job)",
+        "hide": 0,
+        "includeAll": false,
+        "label": "Job",
+        "multi": false,
+        "name": "job",
+        "options": [],
+        "query": {
+          "query": "label_values(nvidia_smi_index,job)",
+          "refId": "Prometheus-job-Variable-Query"
+        },
+        "refresh": 1,
+        "regex": "",
+        "skipUrlSync": false,
+        "tagValuesQuery": "",
+        "tagsQuery": "",
+        "type": "query",
+        "useTags": false
+      },
+      {
+        "current": {},
+        "datasource": {
+          "type": "prometheus",
+          "uid": "${DS_PROMETHEUS}"
+        },
+        "definition": "label_values(nvidia_smi_index{job=\"$job\"},instance)",
+        "hide": 0,
+        "includeAll": false,
+        "label": "Host",
+        "multi": false,
+        "name": "node",
+        "options": [],
+        "query": {
+          "query": "label_values(nvidia_smi_index{job=\"$job\"},instance)",
+          "refId": "Prometheus-node-Variable-Query"
+        },
+        "refresh": 1,
+        "regex": "",
+        "skipUrlSync": false,
+        "tagValuesQuery": "",
+        "tagsQuery": "",
+        "type": "query",
+        "useTags": false
+      },
+      {
+        "current": {},
+        "datasource": {
+          "type": "prometheus",
+          "uid": "${DS_PROMETHEUS}"
+        },
+        "definition": "label_values(nvidia_smi_index{instance=\"$node\"},uuid)",
+        "hide": 0,
+        "includeAll": false,
+        "label": "GPU",
+        "multi": false,
+        "name": "gpu",
+        "options": [],
+        "query": {
+          "query": "label_values(nvidia_smi_index{instance=\"$node\"},uuid)",
+          "refId": "StandardVariableQuery"
+        },
+        "refresh": 1,
+        "regex": "",
+        "skipUrlSync": false,
+        "sort": 1,
+        "tagValuesQuery": "",
+        "tagsQuery": "",
+        "type": "query",
+        "useTags": false
+      }
+    ]
+  },
+  "time": {
+    "from": "now-30m",
+    "to": "now"
+  },
+  "timepicker": {},
+  "timezone": "",
+  "title": "Nvidia GPU Metrics",
+  "uid": "vlvPlrgnk",
+  "version": 8,
+  "weekStart": ""
+}
\ No newline at end of file
diff --git a/ansible/playbooks/saas/roles/grafana/files/dashboards/vllm/official-dashboard.json b/ansible/playbooks/saas/roles/grafana/files/dashboards/vllm/official-dashboard.json
new file mode 100644
index 00000000..35b921de
--- /dev/null
+++ b/ansible/playbooks/saas/roles/grafana/files/dashboards/vllm/official-dashboard.json
@@ -0,0 +1,1353 @@
+{
+  "annotations": {
+    "list": [
+      {
+        "builtIn": 1,
+        "datasource": {
+          "type": "grafana",
+          "uid": "-- Grafana --"
+        },
+        "enable": true,
+        "hide": true,
+        "iconColor": "rgba(0,211,255,1)",
+        "name": "Annotations & Alerts",
+        "type": "dashboard"
+      }
+    ]
+  },
+  "editable": true,
+  "fiscalYearStartMonth": 0,
+  "graphTooltip": 0,
+  "id": 0,
+  "links": [],
+  "panels": [
+    {
+      "collapsed": false,
+      "gridPos": {
+        "h": 1,
+        "w": 24,
+        "x": 0,
+        "y": 0
+      },
+      "id": 100,
+      "panels": [],
+      "title": "Overview System Performance",
+      "type": "row"
+    },
+    {
+      "datasource": {
+        "type": "prometheus",
+        "uid": "prometheus"
+      },
+      "description": "Number of healthy vLLM instances (by instance usage)",
+      "fieldConfig": {
+        "defaults": {
+          "color": {
+            "mode": "thresholds"
+          },
+          "mappings": [],
+          "thresholds": {
+            "mode": "absolute",
+            "steps": [
+              {
+                "color": "green",
+                "value": 0
+              },
+              {
+                "color": "red",
+                "value": 80
+              }
+            ]
+          }
+        },
+        "overrides": []
+      },
+      "gridPos": {
+        "h": 7,
+        "w": 12,
+        "x": 0,
+        "y": 1
+      },
+      "id": 1,
+      "interval": "60s",
+      "options": {
+        "colorMode": "value",
+        "graphMode": "area",
+        "justifyMode": "auto",
+        "orientation": "auto",
+        "percentChangeColorMode": "standard",
+        "reduceOptions": {
+          "calcs": [
+            "lastNotNull"
+          ],
+          "fields": "",
+          "values": false
+        },
+        "showPercentChange": false,
+        "textMode": "auto",
+        "wideLayout": true
+      },
+      "pluginVersion": "12.2.0",
+      "targets": [
+        {
+          "disableTextWrap": false,
+          "editorMode": "builder",
+          "expr": "count by(endpoint) (vllm:healthy_pods_total)",
+          "format": "time_series",
+          "fullMetaSearch": false,
+          "includeNullMetadata": true,
+          "legendFormat": "vLLM instances",
+          "range": true,
+          "refId": "A",
+          "useBackend": false
+        }
+      ],
+      "title": "Available vLLM instances",
+      "type": "stat"
+    },
+    {
+      "datasource": {
+        "type": "prometheus",
+        "uid": "prometheus"
+      },
+      "description": "Average end-to-end request latency in seconds",
+      "fieldConfig": {
+        "defaults": {
+          "color": {
+            "mode": "thresholds"
+          },
+          "mappings": [],
+          "thresholds": {
+            "mode": "absolute",
+            "steps": [
+              {
+                "color": "green",
+                "value": 0
+              },
+              {
+                "color": "red",
+                "value": 1
+              }
+            ]
+          }
+        },
+        "overrides": []
+      },
+      "gridPos": {
+        "h": 7,
+        "w": 12,
+        "x": 12,
+        "y": 1
+      },
+      "id": 19,
+      "interval": "60s",
+      "options": {
+        "colorMode": "value",
+        "graphMode": "area",
+        "justifyMode": "auto",
+        "orientation": "auto",
+        "percentChangeColorMode": "standard",
+        "reduceOptions": {
+          "calcs": [
+            "lastNotNull"
+          ],
+          "fields": "",
+          "values": false
+        },
+        "showPercentChange": false,
+        "textMode": "auto",
+        "wideLayout": true
+      },
+      "pluginVersion": "12.2.0",
+      "targets": [
+        {
+          "editorMode": "builder",
+          "expr": "avg(vllm:e2e_request_latency_seconds_sum) / avg(vllm:e2e_request_latency_seconds_count)",
+          "legendFormat": "Avg Latency",
+          "range": true,
+          "refId": "A"
+        }
+      ],
+      "title": "Average Latency",
+      "type": "stat"
+    },
+    {
+      "datasource": {
+        "type": "prometheus",
+        "uid": "prometheus"
+      },
+      "fieldConfig": {
+        "defaults": {
+          "color": {
+            "mode": "thresholds"
+          },
+          "mappings": [],
+          "thresholds": {
+            "mode": "absolute",
+            "steps": [
+              {
+                "color": "green",
+                "value": 0
+              },
+              {
+                "color": "red",
+                "value": 80
+              }
+            ]
+          }
+        },
+        "overrides": []
+      },
+      "gridPos": {
+        "h": 7,
+        "w": 24,
+        "x": 0,
+        "y": 8
+      },
+      "id": 2,
+      "interval": "60s",
+      "options": {
+        "displayMode": "gradient",
+        "legend": {
+          "calcs": [],
+          "displayMode": "list",
+          "placement": "bottom",
+          "showLegend": false
+        },
+        "maxVizHeight": 300,
+        "minVizHeight": 16,
+        "minVizWidth": 8,
+        "namePlacement": "auto",
+        "orientation": "auto",
+        "reduceOptions": {
+          "calcs": [
+            "lastNotNull"
+          ],
+          "fields": "",
+          "values": false
+        },
+        "showUnfilled": true,
+        "sizing": "auto",
+        "valueMode": "color"
+      },
+      "pluginVersion": "12.2.0",
+      "targets": [
+        {
+          "editorMode": "builder",
+          "expr": "sum by(le) (vllm:e2e_request_latency_seconds_bucket)",
+          "format": "heatmap",
+          "legendFormat": "{{le}}",
+          "range": true,
+          "refId": "A"
+        }
+      ],
+      "title": "Request latency distribution",
+      "type": "bargauge"
+    },
+    {
+      "collapsed": false,
+      "gridPos": {
+        "h": 1,
+        "w": 24,
+        "x": 0,
+        "y": 15
+      },
+      "id": 101,
+      "panels": [],
+      "title": "QoS Information",
+      "type": "row"
+    },
+    {
+      "datasource": {
+        "type": "prometheus",
+        "uid": "prometheus"
+      },
+      "fieldConfig": {
+        "defaults": {
+          "color": {
+            "mode": "thresholds"
+          },
+          "mappings": [],
+          "thresholds": {
+            "mode": "absolute",
+            "steps": [
+              {
+                "color": "green",
+                "value": 0
+              },
+              {
+                "color": "red",
+                "value": 80
+              }
+            ]
+          }
+        },
+        "overrides": []
+      },
+      "gridPos": {
+        "h": 7,
+        "w": 12,
+        "x": 0,
+        "y": 16
+      },
+      "id": 3,
+      "interval": "60s",
+      "options": {
+        "colorMode": "value",
+        "graphMode": "area",
+        "justifyMode": "auto",
+        "orientation": "auto",
+        "percentChangeColorMode": "standard",
+        "reduceOptions": {
+          "calcs": [
+            "lastNotNull"
+          ],
+          "fields": "",
+          "values": false
+        },
+        "showPercentChange": false,
+        "textMode": "auto",
+        "wideLayout": true
+      },
+      "pluginVersion": "12.2.0",
+      "targets": [
+        {
+          "editorMode": "builder",
+          "expr": "vllm:current_qps",
+          "legendFormat": "Current QPS",
+          "range": true,
+          "refId": "A"
+        }
+      ],
+      "title": "Current QPS",
+      "type": "stat"
+    },
+    {
+      "datasource": {
+        "type": "prometheus",
+        "uid": "prometheus"
+      },
+      "description": "Average Inter-Token Latency",
+      "fieldConfig": {
+        "defaults": {
+          "color": {
+            "mode": "thresholds"
+          },
+          "mappings": [],
+          "thresholds": {
+            "mode": "absolute",
+            "steps": [
+              {
+                "color": "green",
+                "value": 0
+              },
+              {
+                "color": "red",
+                "value": 1
+              }
+            ]
+          }
+        },
+        "overrides": []
+      },
+      "gridPos": {
+        "h": 7,
+        "w": 12,
+        "x": 12,
+        "y": 16
+      },
+      "id": 20,
+      "interval": "60s",
+      "options": {
+        "colorMode": "value",
+        "graphMode": "area",
+        "justifyMode": "auto",
+        "orientation": "auto",
+        "percentChangeColorMode": "standard",
+        "reduceOptions": {
+          "calcs": [
+            "lastNotNull"
+          ],
+          "fields": "",
+          "values": false
+        },
+        "showPercentChange": false,
+        "textMode": "auto",
+        "wideLayout": true
+      },
+      "pluginVersion": "12.2.0",
+      "targets": [
+        {
+          "editorMode": "builder",
+          "expr": "avg(vllm:time_per_output_token_seconds_sum) / avg(vllm:time_per_output_token_seconds_count)",
+          "legendFormat": "Avg ITL",
+          "range": true,
+          "refId": "A"
+        }
+      ],
+      "title": "Average ITL",
+      "type": "stat"
+    },
+    {
+      "datasource": {
+        "type": "prometheus",
+        "uid": "prometheus"
+      },
+      "fieldConfig": {
+        "defaults": {
+          "color": {
+            "mode": "thresholds"
+          },
+          "mappings": [],
+          "thresholds": {
+            "mode": "absolute",
+            "steps": [
+              {
+                "color": "green",
+                "value": 0
+              },
+              {
+                "color": "red",
+                "value": 80
+              }
+            ]
+          }
+        },
+        "overrides": []
+      },
+      "gridPos": {
+        "h": 7,
+        "w": 24,
+        "x": 0,
+        "y": 23
+      },
+      "id": 4,
+      "interval": "60s",
+      "options": {
+        "displayMode": "gradient",
+        "legend": {
+          "calcs": [],
+          "displayMode": "list",
+          "placement": "bottom",
+          "showLegend": false
+        },
+        "maxVizHeight": 300,
+        "minVizHeight": 16,
+        "minVizWidth": 8,
+        "namePlacement": "auto",
+        "orientation": "auto",
+        "reduceOptions": {
+          "calcs": [
+            "lastNotNull"
+          ],
+          "fields": "",
+          "values": false
+        },
+        "showUnfilled": true,
+        "sizing": "auto",
+        "valueMode": "color"
+      },
+      "pluginVersion": "12.2.0",
+      "targets": [
+        {
+          "editorMode": "builder",
+          "expr": "sum by(le) (vllm:time_to_first_token_seconds_bucket)",
+          "format": "heatmap",
+          "legendFormat": "__auto",
+          "range": true,
+          "refId": "A"
+        }
+      ],
+      "title": "Request TTFT distribution",
+      "type": "bargauge"
+    },
+    {
+      "collapsed": false,
+      "gridPos": {
+        "h": 1,
+        "w": 24,
+        "x": 0,
+        "y": 30
+      },
+      "id": 102,
+      "panels": [],
+      "title": "Serving Engine Load",
+      "type": "row"
+    },
+    {
+      "datasource": {
+        "type": "prometheus",
+        "uid": "prometheus"
+      },
+      "fieldConfig": {
+        "defaults": {
+          "color": {
+            "mode": "palette-classic"
+          },
+          "custom": {
+            "axisBorderShow": false,
+            "axisCenteredZero": false,
+            "axisColorMode": "text",
+            "axisLabel": "",
+            "axisPlacement": "auto",
+            "barAlignment": 0,
+            "barWidthFactor": 0.6,
+            "drawStyle": "line",
+            "fillOpacity": 0,
+            "gradientMode": "none",
+            "hideFrom": {
+              "legend": false,
+              "tooltip": false,
+              "viz": false
+            },
+            "insertNulls": false,
+            "lineInterpolation": "linear",
+            "lineWidth": 1,
+            "pointSize": 5,
+            "scaleDistribution": {
+              "type": "linear"
+            },
+            "showPoints": "auto",
+            "showValues": false,
+            "spanNulls": false,
+            "stacking": {
+              "group": "A",
+              "mode": "none"
+            },
+            "thresholdsStyle": {
+              "mode": "off"
+            }
+          },
+          "mappings": [],
+          "thresholds": {
+            "mode": "absolute",
+            "steps": [
+              {
+                "color": "green",
+                "value": 0
+              },
+              {
+                "color": "red",
+                "value": 80
+              }
+            ]
+          }
+        },
+        "overrides": []
+      },
+      "gridPos": {
+        "h": 8,
+        "w": 8,
+        "x": 0,
+        "y": 31
+      },
+      "id": 10,
+      "interval": "60s",
+      "options": {
+        "legend": {
+          "calcs": [],
+          "displayMode": "list",
+          "placement": "right",
+          "showLegend": true
+        },
+        "tooltip": {
+          "hideZeros": false,
+          "mode": "single",
+          "sort": "none"
+        }
+      },
+      "pluginVersion": "12.2.0",
+      "targets": [
+        {
+          "editorMode": "code",
+          "expr": "vllm:num_requests_running",
+          "legendFormat": "{{instance}}",
+          "range": true,
+          "refId": "A"
+        }
+      ],
+      "title": "Number of Running Requests",
+      "type": "timeseries"
+    },
+    {
+      "datasource": {
+        "type": "prometheus",
+        "uid": "prometheus"
+      },
+      "fieldConfig": {
+        "defaults": {
+          "color": {
+            "mode": "palette-classic"
+          },
+          "custom": {
+            "axisBorderShow": false,
+            "axisCenteredZero": false,
+            "axisColorMode": "text",
+            "axisLabel": "",
+            "axisPlacement": "auto",
+            "barAlignment": 0,
+            "barWidthFactor": 0.6,
+            "drawStyle": "line",
+            "fillOpacity": 0,
+            "gradientMode": "none",
+            "hideFrom": {
+              "legend": false,
+              "tooltip": false,
+              "viz": false
+            },
+            "insertNulls": false,
+            "lineInterpolation": "linear",
+            "lineWidth": 1,
+            "pointSize": 5,
+            "scaleDistribution": {
+              "type": "linear"
+            },
+            "showPoints": "auto",
+            "showValues": false,
+            "spanNulls": false,
+            "stacking": {
+              "group": "A",
+              "mode": "none"
+            },
+            "thresholdsStyle": {
+              "mode": "off"
+            }
+          },
+          "mappings": [],
+          "thresholds": {
+            "mode": "absolute",
+            "steps": [
+              {
+                "color": "green",
+                "value": 0
+              },
+              {
+                "color": "red",
+                "value": 80
+              }
+            ]
+          }
+        },
+        "overrides": []
+      },
+      "gridPos": {
+        "h": 8,
+        "w": 8,
+        "x": 8,
+        "y": 31
+      },
+      "id": 11,
+      "interval": "60s",
+      "options": {
+        "legend": {
+          "calcs": [],
+          "displayMode": "list",
+          "placement": "right",
+          "showLegend": true
+        },
+        "tooltip": {
+          "hideZeros": false,
+          "mode": "single",
+          "sort": "none"
+        }
+      },
+      "pluginVersion": "12.2.0",
+      "targets": [
+        {
+          "editorMode": "builder",
+          "expr": "vllm:num_requests_waiting",
+          "legendFormat": "{{instance}}",
+          "range": true,
+          "refId": "A"
+        }
+      ],
+      "title": "Number of Pending Requests",
+      "type": "timeseries"
+    },
+    {
+      "datasource": {
+        "type": "prometheus",
+        "uid": "prometheus"
+      },
+      "fieldConfig": {
+        "defaults": {
+          "color": {
+            "mode": "palette-classic"
+          },
+          "custom": {
+            "axisBorderShow": false,
+            "axisCenteredZero": false,
+            "axisColorMode": "text",
+            "axisLabel": "",
+            "axisPlacement": "auto",
+            "barAlignment": 0,
+            "barWidthFactor": 0.6,
+            "drawStyle": "line",
+            "fillOpacity": 0,
+            "gradientMode": "none",
+            "hideFrom": {
+              "legend": false,
+              "tooltip": false,
+              "viz": false
+            },
+            "insertNulls": false,
+            "lineInterpolation": "linear",
+            "lineWidth": 1,
+            "pointSize": 5,
+            "scaleDistribution": {
+              "type": "linear"
+            },
+            "showPoints": "auto",
+            "showValues": false,
+            "spanNulls": false,
+            "stacking": {
+              "group": "A",
+              "mode": "none"
+            },
+            "thresholdsStyle": {
+              "mode": "off"
+            }
+          },
+          "mappings": [],
+          "thresholds": {
+            "mode": "absolute",
+            "steps": [
+              {
+                "color": "green",
+                "value": 0
+              },
+              {
+                "color": "red",
+                "value": 80
+              }
+            ]
+          }
+        },
+        "overrides": []
+      },
+      "gridPos": {
+        "h": 8,
+        "w": 8,
+        "x": 16,
+        "y": 31
+      },
+      "id": 12,
+      "interval": "60s",
+      "options": {
+        "legend": {
+          "calcs": [],
+          "displayMode": "list",
+          "placement": "right",
+          "showLegend": true
+        },
+        "tooltip": {
+          "hideZeros": false,
+          "mode": "single",
+          "sort": "none"
+        }
+      },
+      "pluginVersion": "12.2.0",
+      "targets": [
+        {
+          "editorMode": "builder",
+          "expr": "vllm:gpu_cache_usage_perc",
+          "legendFormat": "{{instance}}",
+          "range": true,
+          "refId": "A"
+        }
+      ],
+      "title": "GPU KV Usage Percentage",
+      "type": "timeseries"
+    },
+    {
+      "datasource": {
+        "type": "prometheus",
+        "uid": "prometheus"
+      },
+      "fieldConfig": {
+        "defaults": {
+          "color": {
+            "mode": "palette-classic"
+          },
+          "custom": {
+            "axisBorderShow": false,
+            "axisCenteredZero": false,
+            "axisColorMode": "text",
+            "axisLabel": "",
+            "axisPlacement": "auto",
+            "barAlignment": 0,
+            "barWidthFactor": 0.6,
+            "drawStyle": "line",
+            "fillOpacity": 0,
+            "gradientMode": "none",
+            "hideFrom": {
+              "legend": false,
+              "tooltip": false,
+              "viz": false
+            },
+            "insertNulls": false,
+            "lineInterpolation": "linear",
+            "lineWidth": 1,
+            "pointSize": 5,
+            "scaleDistribution": {
+              "type": "linear"
+            },
+            "showPoints": "auto",
+            "showValues": false,
+            "spanNulls": false,
+            "stacking": {
+              "group": "A",
+              "mode": "none"
+            },
+            "thresholdsStyle": {
+              "mode": "off"
+            }
+          },
+          "mappings": [],
+          "thresholds": {
+            "mode": "absolute",
+            "steps": [
+              {
+                "color": "green",
+                "value": 0
+              },
+              {
+                "color": "red",
+                "value": 80
+              }
+            ]
+          }
+        },
+        "overrides": []
+      },
+      "gridPos": {
+        "h": 8,
+        "w": 12,
+        "x": 0,
+        "y": 39
+      },
+      "id": 13,
+      "interval": "60s",
+      "options": {
+        "legend": {
+          "calcs": [],
+          "displayMode": "list",
+          "placement": "right",
+          "showLegend": true
+        },
+        "tooltip": {
+          "hideZeros": false,
+          "mode": "single",
+          "sort": "none"
+        }
+      },
+      "pluginVersion": "12.2.0",
+      "targets": [
+        {
+          "disableTextWrap": false,
+          "editorMode": "builder",
+          "expr": "vllm:gpu_prefix_cache_hits_total{endpoint=\"service-port\"} / vllm:gpu_prefix_cache_queries_total",
+          "fullMetaSearch": false,
+          "includeNullMetadata": true,
+          "legendFormat": "{{instance}}",
+          "range": true,
+          "refId": "A",
+          "useBackend": false
+        }
+      ],
+      "title": "GPU KV Cache Hit Rate",
+      "type": "timeseries"
+    },
+    {
+      "datasource": {
+        "type": "prometheus",
+        "uid": "prometheus"
+      },
+      "description": "Requests moved from GPU to CPU",
+      "fieldConfig": {
+        "defaults": {
+          "color": {
+            "mode": "thresholds"
+          },
+          "mappings": [],
+          "thresholds": {
+            "mode": "absolute",
+            "steps": [
+              {
+                "color": "green",
+                "value": 0
+              },
+              {
+                "color": "red",
+                "value": 0
+              }
+            ]
+          }
+        },
+        "overrides": []
+      },
+      "gridPos": {
+        "h": 8,
+        "w": 12,
+        "x": 12,
+        "y": 39
+      },
+      "id": 21,
+      "interval": "60s",
+      "options": {
+        "colorMode": "value",
+        "graphMode": "area",
+        "justifyMode": "auto",
+        "orientation": "auto",
+        "percentChangeColorMode": "standard",
+        "reduceOptions": {
+          "calcs": [
+            "lastNotNull"
+          ],
+          "fields": "",
+          "values": false
+        },
+        "showPercentChange": false,
+        "textMode": "auto",
+        "wideLayout": true
+      },
+      "pluginVersion": "12.2.0",
+      "targets": [
+        {
+          "editorMode": "builder",
+          "expr": "vllm:num_requests_swapped",
+          "legendFormat": "Swapped Requests",
+          "range": true,
+          "refId": "A"
+        }
+      ],
+      "title": "Number of Swapped Requests",
+      "type": "stat"
+    },
+    {
+      "collapsed": false,
+      "gridPos": {
+        "h": 1,
+        "w": 24,
+        "x": 0,
+        "y": 47
+      },
+      "id": 103,
+      "panels": [],
+      "title": "Current Resource Usage",
+      "type": "row"
+    },
+    {
+      "datasource": {
+        "type": "prometheus",
+        "uid": "prometheus"
+      },
+      "fieldConfig": {
+        "defaults": {
+          "color": {
+            "mode": "palette-classic"
+          },
+          "custom": {
+            "axisBorderShow": false,
+            "axisCenteredZero": false,
+            "axisColorMode": "text",
+            "axisLabel": "",
+            "axisPlacement": "auto",
+            "barAlignment": 0,
+            "barWidthFactor": 0.6,
+            "drawStyle": "line",
+            "fillOpacity": 0,
+            "gradientMode": "none",
+            "hideFrom": {
+              "legend": false,
+              "tooltip": false,
+              "viz": false
+            },
+            "insertNulls": false,
+            "lineInterpolation": "linear",
+            "lineWidth": 1,
+            "pointSize": 5,
+            "scaleDistribution": {
+              "type": "linear"
+            },
+            "showPoints": "auto",
+            "showValues": false,
+            "spanNulls": false,
+            "stacking": {
+              "group": "A",
+              "mode": "none"
+            },
+            "thresholdsStyle": {
+              "mode": "off"
+            }
+          },
+          "mappings": [],
+          "thresholds": {
+            "mode": "absolute",
+            "steps": [
+              {
+                "color": "green",
+                "value": 0
+              },
+              {
+                "color": "red",
+                "value": 80
+              }
+            ]
+          }
+        },
+        "overrides": []
+      },
+      "gridPos": {
+        "h": 8,
+        "w": 6,
+        "x": 0,
+        "y": 48
+      },
+      "id": 14,
+      "interval": "60s",
+      "options": {
+        "legend": {
+          "calcs": [],
+          "displayMode": "list",
+          "placement": "right",
+          "showLegend": true
+        },
+        "tooltip": {
+          "hideZeros": false,
+          "mode": "single",
+          "sort": "none"
+        }
+      },
+      "pluginVersion": "12.2.0",
+      "targets": [
+        {
+          "disableTextWrap": false,
+          "editorMode": "builder",
+          "expr": "vllm:gpu_cache_usage_perc",
+          "fullMetaSearch": false,
+          "includeNullMetadata": true,
+          "legendFormat": "GPU Usage",
+          "range": true,
+          "refId": "A",
+          "useBackend": false
+        }
+      ],
+      "title": "GPU Cache Usage (%)",
+      "type": "timeseries"
+    },
+    {
+      "datasource": {
+        "type": "prometheus",
+        "uid": "prometheus"
+      },
+      "fieldConfig": {
+        "defaults": {
+          "color": {
+            "mode": "palette-classic"
+          },
+          "custom": {
+            "axisBorderShow": false,
+            "axisCenteredZero": false,
+            "axisColorMode": "text",
+            "axisLabel": "",
+            "axisPlacement": "auto",
+            "barAlignment": 0,
+            "barWidthFactor": 0.6,
+            "drawStyle": "line",
+            "fillOpacity": 0,
+            "gradientMode": "none",
+            "hideFrom": {
+              "legend": false,
+              "tooltip": false,
+              "viz": false
+            },
+            "insertNulls": false,
+            "lineInterpolation": "linear",
+            "lineWidth": 1,
+            "pointSize": 5,
+            "scaleDistribution": {
+              "type": "linear"
+            },
+            "showPoints": "auto",
+            "showValues": false,
+            "spanNulls": false,
+            "stacking": {
+              "group": "A",
+              "mode": "none"
+            },
+            "thresholdsStyle": {
+              "mode": "off"
+            }
+          },
+          "mappings": [],
+          "thresholds": {
+            "mode": "absolute",
+            "steps": [
+              {
+                "color": "green",
+                "value": 0
+              },
+              {
+                "color": "red",
+                "value": 80
+              }
+            ]
+          }
+        },
+        "overrides": []
+      },
+      "gridPos": {
+        "h": 8,
+        "w": 6,
+        "x": 6,
+        "y": 48
+      },
+      "id": 15,
+      "interval": "60s",
+      "options": {
+        "legend": {
+          "calcs": [],
+          "displayMode": "list",
+          "placement": "right",
+          "showLegend": true
+        },
+        "tooltip": {
+          "hideZeros": false,
+          "mode": "single",
+          "sort": "none"
+        }
+      },
+      "pluginVersion": "12.2.0",
+      "targets": [
+        {
+          "disableTextWrap": false,
+          "editorMode": "builder",
+          "expr": "router_cpu_usage_percent",
+          "fullMetaSearch": false,
+          "includeNullMetadata": true,
+          "legendFormat": "CPU Usage",
+          "range": true,
+          "refId": "A",
+          "useBackend": false
+        }
+      ],
+      "title": "CPU Usage (%)",
+      "type": "timeseries"
+    },
+    {
+      "datasource": {
+        "type": "prometheus",
+        "uid": "prometheus"
+      },
+      "fieldConfig": {
+        "defaults": {
+          "color": {
+            "mode": "palette-classic"
+          },
+          "custom": {
+            "axisBorderShow": false,
+            "axisCenteredZero": false,
+            "axisColorMode": "text",
+            "axisLabel": "",
+            "axisPlacement": "auto",
+            "barAlignment": 0,
+            "barWidthFactor": 0.6,
+            "drawStyle": "line",
+            "fillOpacity": 0,
+            "gradientMode": "none",
+            "hideFrom": {
+              "legend": false,
+              "tooltip": false,
+              "viz": false
+            },
+            "insertNulls": false,
+            "lineInterpolation": "linear",
+            "lineWidth": 1,
+            "pointSize": 5,
+            "scaleDistribution": {
+              "type": "linear"
+            },
+            "showPoints": "auto",
+            "showValues": false,
+            "spanNulls": false,
+            "stacking": {
+              "group": "A",
+              "mode": "none"
+            },
+            "thresholdsStyle": {
+              "mode": "off"
+            }
+          },
+          "mappings": [],
+          "thresholds": {
+            "mode": "absolute",
+            "steps": [
+              {
+                "color": "green",
+                "value": 0
+              },
+              {
+                "color": "red",
+                "value": 80
+              }
+            ]
+          }
+        },
+        "overrides": []
+      },
+      "gridPos": {
+        "h": 8,
+        "w": 6,
+        "x": 12,
+        "y": 48
+      },
+      "id": 16,
+      "interval": "60s",
+      "options": {
+        "legend": {
+          "calcs": [],
+          "displayMode": "list",
+          "placement": "right",
+          "showLegend": true
+        },
+        "tooltip": {
+          "hideZeros": false,
+          "mode": "single",
+          "sort": "none"
+        }
+      },
+      "pluginVersion": "12.2.0",
+      "targets": [
+        {
+          "disableTextWrap": false,
+          "editorMode": "builder",
+          "expr": "router_memory_usage_percent",
+          "fullMetaSearch": false,
+          "includeNullMetadata": true,
+          "legendFormat": "Memory Usage",
+          "range": true,
+          "refId": "A",
+          "useBackend": false
+        }
+      ],
+      "title": "Memory Usage (%)",
+      "type": "timeseries"
+    },
+    {
+      "datasource": {
+        "type": "prometheus",
+        "uid": "prometheus"
+      },
+      "fieldConfig": {
+        "defaults": {
+          "color": {
+            "mode": "palette-classic"
+          },
+          "custom": {
+            "axisBorderShow": false,
+            "axisCenteredZero": false,
+            "axisColorMode": "text",
+            "axisLabel": "",
+            "axisPlacement": "auto",
+            "barAlignment": 0,
+            "barWidthFactor": 0.6,
+            "drawStyle": "line",
+            "fillOpacity": 0,
+            "gradientMode": "none",
+            "hideFrom": {
+              "legend": false,
+              "tooltip": false,
+              "viz": false
+            },
+            "insertNulls": false,
+            "lineInterpolation": "linear",
+            "lineWidth": 1,
+            "pointSize": 5,
+            "scaleDistribution": {
+              "type": "linear"
+            },
+            "showPoints": "auto",
+            "showValues": false,
+            "spanNulls": false,
+            "stacking": {
+              "group": "A",
+              "mode": "none"
+            },
+            "thresholdsStyle": {
+              "mode": "off"
+            }
+          },
+          "mappings": [],
+          "thresholds": {
+            "mode": "absolute",
+            "steps": [
+              {
+                "color": "green",
+                "value": 0
+              },
+              {
+                "color": "red",
+                "value": 80
+              }
+            ]
+          }
+        },
+        "overrides": []
+      },
+      "gridPos": {
+        "h": 8,
+        "w": 6,
+        "x": 18,
+        "y": 48
+      },
+      "id": 17,
+      "interval": "60s",
+      "options": {
+        "legend": {
+          "calcs": [],
+          "displayMode": "list",
+          "placement": "right",
+          "showLegend": true
+        },
+        "tooltip": {
+          "hideZeros": false,
+          "mode": "single",
+          "sort": "none"
+        }
+      },
+      "pluginVersion": "12.2.0",
+      "targets": [
+        {
+          "disableTextWrap": false,
+          "editorMode": "builder",
+          "expr": "router_disk_usage_percent",
+          "fullMetaSearch": false,
+          "includeNullMetadata": true,
+          "legendFormat": "Disk Usage",
+          "range": true,
+          "refId": "A",
+          "useBackend": false
+        }
+      ],
+      "title": "Disk Usage (%)",
+      "type": "timeseries"
+    }
+  ],
+  "preload": false,
+  "refresh": "auto",
+  "schemaVersion": 42,
+  "tags": [],
+  "templating": {
+    "list": []
+  },
+  "time": {
+    "from": "now-15m",
+    "to": "now"
+  },
+  "timepicker": {},
+  "timezone": "browser",
+  "title": "vLLM Dashboard",
+  "uid": "750918234",
+  "version": 2
+}
\ No newline at end of file
diff --git a/ansible/playbooks/saas/roles/grafana/files/dashboards/vllm/vllm.json b/ansible/playbooks/saas/roles/grafana/files/dashboards/vllm/vllm.json
new file mode 100644
index 00000000..7b5fc131
--- /dev/null
+++ b/ansible/playbooks/saas/roles/grafana/files/dashboards/vllm/vllm.json
@@ -0,0 +1,1561 @@
+{
+  "annotations": {
+    "list": [
+      {
+        "builtIn": 1,
+        "datasource": {
+          "type": "grafana",
+          "uid": "-- Grafana --"
+        },
+        "enable": true,
+        "hide": true,
+        "iconColor": "rgba(0, 211, 255, 1)",
+        "name": "Annotations & Alerts",
+        "target": {
+          "limit": 100,
+          "matchAny": false,
+          "tags": [],
+          "type": "dashboard"
+        },
+        "type": "dashboard"
+      }
+    ]
+  },
+  "description": "Monitoring vLLM Inference Server",
+  "editable": true,
+  "fiscalYearStartMonth": 0,
+  "graphTooltip": 0,
+  "id": 0,
+  "links": [],
+  "panels": [
+    {
+      "datasource": {
+        "type": "prometheus",
+        "uid": "${DS_PROMETHEUS}"
+      },
+      "description": "End to end request latency measured in seconds.",
+      "fieldConfig": {
+        "defaults": {
+          "color": {
+            "mode": "palette-classic"
+          },
+          "custom": {
+            "axisBorderShow": false,
+            "axisCenteredZero": false,
+            "axisColorMode": "text",
+            "axisLabel": "",
+            "axisPlacement": "auto",
+            "barAlignment": 0,
+            "barWidthFactor": 0.6,
+            "drawStyle": "line",
+            "fillOpacity": 0,
+            "gradientMode": "none",
+            "hideFrom": {
+              "legend": false,
+              "tooltip": false,
+              "viz": false
+            },
+            "insertNulls": false,
+            "lineInterpolation": "linear",
+            "lineWidth": 1,
+            "pointSize": 5,
+            "scaleDistribution": {
+              "type": "linear"
+            },
+            "showPoints": "auto",
+            "showValues": false,
+            "spanNulls": false,
+            "stacking": {
+              "group": "A",
+              "mode": "none"
+            },
+            "thresholdsStyle": {
+              "mode": "off"
+            }
+          },
+          "mappings": [],
+          "thresholds": {
+            "mode": "absolute",
+            "steps": [
+              {
+                "color": "green",
+                "value": 0
+              },
+              {
+                "color": "red",
+                "value": 80
+              }
+            ]
+          },
+          "unit": "s"
+        },
+        "overrides": []
+      },
+      "gridPos": {
+        "h": 8,
+        "w": 12,
+        "x": 0,
+        "y": 0
+      },
+      "id": 9,
+      "interval": "60s",
+      "options": {
+        "legend": {
+          "calcs": [],
+          "displayMode": "list",
+          "placement": "bottom",
+          "showLegend": true
+        },
+        "tooltip": {
+          "hideZeros": false,
+          "mode": "single",
+          "sort": "none"
+        }
+      },
+      "pluginVersion": "12.2.0",
+      "targets": [
+        {
+          "datasource": {
+            "type": "prometheus",
+            "uid": "${DS_PROMETHEUS}"
+          },
+          "disableTextWrap": false,
+          "editorMode": "builder",
+          "expr": "histogram_quantile(0.99, sum by(le) (rate(vllm:e2e_request_latency_seconds_bucket{model_name=\"$model_name\"}[$__rate_interval])))",
+          "fullMetaSearch": false,
+          "includeNullMetadata": false,
+          "instant": false,
+          "legendFormat": "P99",
+          "range": true,
+          "refId": "A",
+          "useBackend": false
+        },
+        {
+          "datasource": {
+            "type": "prometheus",
+            "uid": "${DS_PROMETHEUS}"
+          },
+          "disableTextWrap": false,
+          "editorMode": "builder",
+          "expr": "histogram_quantile(0.95, sum by(le) (rate(vllm:e2e_request_latency_seconds_bucket{model_name=\"$model_name\"}[$__rate_interval])))",
+          "fullMetaSearch": false,
+          "hide": false,
+          "includeNullMetadata": false,
+          "instant": false,
+          "legendFormat": "P95",
+          "range": true,
+          "refId": "B",
+          "useBackend": false
+        },
+        {
+          "datasource": {
+            "type": "prometheus",
+            "uid": "${DS_PROMETHEUS}"
+          },
+          "disableTextWrap": false,
+          "editorMode": "builder",
+          "expr": "histogram_quantile(0.9, sum by(le) (rate(vllm:e2e_request_latency_seconds_bucket{model_name=\"$model_name\"}[$__rate_interval])))",
+          "fullMetaSearch": false,
+          "hide": false,
+          "includeNullMetadata": false,
+          "instant": false,
+          "legendFormat": "P90",
+          "range": true,
+          "refId": "C",
+          "useBackend": false
+        },
+        {
+          "datasource": {
+            "type": "prometheus",
+            "uid": "${DS_PROMETHEUS}"
+          },
+          "disableTextWrap": false,
+          "editorMode": "builder",
+          "expr": "histogram_quantile(0.5, sum by(le) (rate(vllm:e2e_request_latency_seconds_bucket{model_name=\"$model_name\"}[$__rate_interval])))",
+          "fullMetaSearch": false,
+          "hide": false,
+          "includeNullMetadata": false,
+          "instant": false,
+          "legendFormat": "P50",
+          "range": true,
+          "refId": "D",
+          "useBackend": false
+        },
+        {
+          "datasource": {
+            "type": "prometheus",
+            "uid": "${DS_PROMETHEUS}"
+          },
+          "editorMode": "code",
+          "expr": "rate(vllm:e2e_request_latency_seconds_sum{model_name=\"$model_name\"}[$__rate_interval])\n/\nrate(vllm:e2e_request_latency_seconds_count{model_name=\"$model_name\"}[$__rate_interval])",
+          "hide": false,
+          "instant": false,
+          "legendFormat": "Average",
+          "range": true,
+          "refId": "E"
+        }
+      ],
+      "title": "E2E Request Latency",
+      "type": "timeseries"
+    },
+    {
+      "datasource": {
+        "type": "prometheus",
+        "uid": "${DS_PROMETHEUS}"
+      },
+      "description": "Number of tokens processed per second",
+      "fieldConfig": {
+        "defaults": {
+          "color": {
+            "mode": "palette-classic"
+          },
+          "custom": {
+            "axisBorderShow": false,
+            "axisCenteredZero": false,
+            "axisColorMode": "text",
+            "axisLabel": "",
+            "axisPlacement": "auto",
+            "barAlignment": 0,
+            "barWidthFactor": 0.6,
+            "drawStyle": "line",
+            "fillOpacity": 0,
+            "gradientMode": "none",
+            "hideFrom": {
+              "legend": false,
+              "tooltip": false,
+              "viz": false
+            },
+            "insertNulls": false,
+            "lineInterpolation": "linear",
+            "lineWidth": 1,
+            "pointSize": 5,
+            "scaleDistribution": {
+              "type": "linear"
+            },
+            "showPoints": "auto",
+            "showValues": false,
+            "spanNulls": false,
+            "stacking": {
+              "group": "A",
+              "mode": "none"
+            },
+            "thresholdsStyle": {
+              "mode": "off"
+            }
+          },
+          "mappings": [],
+          "thresholds": {
+            "mode": "absolute",
+            "steps": [
+              {
+                "color": "green",
+                "value": 0
+              },
+              {
+                "color": "red",
+                "value": 80
+              }
+            ]
+          }
+        },
+        "overrides": []
+      },
+      "gridPos": {
+        "h": 8,
+        "w": 12,
+        "x": 12,
+        "y": 0
+      },
+      "id": 8,
+      "interval": "60s",
+      "options": {
+        "legend": {
+          "calcs": [],
+          "displayMode": "list",
+          "placement": "bottom",
+          "showLegend": true
+        },
+        "tooltip": {
+          "hideZeros": false,
+          "mode": "single",
+          "sort": "none"
+        }
+      },
+      "pluginVersion": "12.2.0",
+      "targets": [
+        {
+          "datasource": {
+            "type": "prometheus",
+            "uid": "${DS_PROMETHEUS}"
+          },
+          "disableTextWrap": false,
+          "editorMode": "code",
+          "expr": "rate(vllm:prompt_tokens_total{model_name=\"$model_name\"}[$__rate_interval])",
+          "fullMetaSearch": false,
+          "includeNullMetadata": false,
+          "instant": false,
+          "legendFormat": "Prompt Tokens/Sec",
+          "range": true,
+          "refId": "A",
+          "useBackend": false
+        },
+        {
+          "datasource": {
+            "type": "prometheus",
+            "uid": "${DS_PROMETHEUS}"
+          },
+          "disableTextWrap": false,
+          "editorMode": "builder",
+          "expr": "rate(vllm:generation_tokens_total{model_name=\"$model_name\"}[$__rate_interval])",
+          "fullMetaSearch": false,
+          "hide": false,
+          "includeNullMetadata": false,
+          "instant": false,
+          "legendFormat": "Generation Tokens/Sec",
+          "range": true,
+          "refId": "B",
+          "useBackend": false
+        }
+      ],
+      "title": "Token Throughput",
+      "type": "timeseries"
+    },
+    {
+      "datasource": {
+        "type": "prometheus",
+        "uid": "${DS_PROMETHEUS}"
+      },
+      "description": "Inter token latency in seconds.",
+      "fieldConfig": {
+        "defaults": {
+          "color": {
+            "mode": "palette-classic"
+          },
+          "custom": {
+            "axisBorderShow": false,
+            "axisCenteredZero": false,
+            "axisColorMode": "text",
+            "axisLabel": "",
+            "axisPlacement": "auto",
+            "barAlignment": 0,
+            "barWidthFactor": 0.6,
+            "drawStyle": "line",
+            "fillOpacity": 0,
+            "gradientMode": "none",
+            "hideFrom": {
+              "legend": false,
+              "tooltip": false,
+              "viz": false
+            },
+            "insertNulls": false,
+            "lineInterpolation": "linear",
+            "lineWidth": 1,
+            "pointSize": 5,
+            "scaleDistribution": {
+              "type": "linear"
+            },
+            "showPoints": "auto",
+            "showValues": false,
+            "spanNulls": false,
+            "stacking": {
+              "group": "A",
+              "mode": "none"
+            },
+            "thresholdsStyle": {
+              "mode": "off"
+            }
+          },
+          "mappings": [],
+          "thresholds": {
+            "mode": "absolute",
+            "steps": [
+              {
+                "color": "green",
+                "value": 0
+              },
+              {
+                "color": "red",
+                "value": 80
+              }
+            ]
+          },
+          "unit": "s"
+        },
+        "overrides": []
+      },
+      "gridPos": {
+        "h": 8,
+        "w": 12,
+        "x": 0,
+        "y": 8
+      },
+      "id": 10,
+      "interval": "60s",
+      "options": {
+        "legend": {
+          "calcs": [],
+          "displayMode": "list",
+          "placement": "bottom",
+          "showLegend": true
+        },
+        "tooltip": {
+          "hideZeros": false,
+          "mode": "single",
+          "sort": "none"
+        }
+      },
+      "pluginVersion": "12.2.0",
+      "targets": [
+        {
+          "datasource": {
+            "type": "prometheus",
+            "uid": "${DS_PROMETHEUS}"
+          },
+          "disableTextWrap": false,
+          "editorMode": "builder",
+          "expr": "histogram_quantile(0.99, sum by(le) (rate(vllm:inter_token_latency_seconds_bucket{model_name=\"$model_name\"}[$__rate_interval])))",
+          "fullMetaSearch": false,
+          "includeNullMetadata": false,
+          "instant": false,
+          "legendFormat": "P99",
+          "range": true,
+          "refId": "A",
+          "useBackend": false
+        },
+        {
+          "datasource": {
+            "type": "prometheus",
+            "uid": "${DS_PROMETHEUS}"
+          },
+          "disableTextWrap": false,
+          "editorMode": "builder",
+          "expr": "histogram_quantile(0.95, sum by(le) (rate(vllm:inter_token_latency_seconds_bucket{model_name=\"$model_name\"}[$__rate_interval])))",
+          "fullMetaSearch": false,
+          "hide": false,
+          "includeNullMetadata": false,
+          "instant": false,
+          "legendFormat": "P95",
+          "range": true,
+          "refId": "B",
+          "useBackend": false
+        },
+        {
+          "datasource": {
+            "type": "prometheus",
+            "uid": "${DS_PROMETHEUS}"
+          },
+          "disableTextWrap": false,
+          "editorMode": "builder",
+          "expr": "histogram_quantile(0.9, sum by(le) (rate(vllm:inter_token_latency_seconds_bucket{model_name=\"$model_name\"}[$__rate_interval])))",
+          "fullMetaSearch": false,
+          "hide": false,
+          "includeNullMetadata": false,
+          "instant": false,
+          "legendFormat": "P90",
+          "range": true,
+          "refId": "C",
+          "useBackend": false
+        },
+        {
+          "datasource": {
+            "type": "prometheus",
+            "uid": "${DS_PROMETHEUS}"
+          },
+          "disableTextWrap": false,
+          "editorMode": "builder",
+          "expr": "histogram_quantile(0.5, sum by(le) (rate(vllm:inter_token_latency_seconds_bucket{model_name=\"$model_name\"}[$__rate_interval])))",
+          "fullMetaSearch": false,
+          "hide": false,
+          "includeNullMetadata": false,
+          "instant": false,
+          "legendFormat": "P50",
+          "range": true,
+          "refId": "D",
+          "useBackend": false
+        },
+        {
+          "datasource": {
+            "type": "prometheus",
+            "uid": "${DS_PROMETHEUS}"
+          },
+          "editorMode": "code",
+          "expr": "rate(vllm:inter_token_latency_seconds_sum{model_name=\"$model_name\"}[$__rate_interval])\n/\nrate(vllm:inter_token_latency_seconds_count{model_name=\"$model_name\"}[$__rate_interval])",
+          "hide": false,
+          "instant": false,
+          "legendFormat": "Mean",
+          "range": true,
+          "refId": "E"
+        }
+      ],
+      "title": "Inter Token Latency",
+      "type": "timeseries"
+    },
+    {
+      "datasource": {
+        "type": "prometheus",
+        "uid": "${DS_PROMETHEUS}"
+      },
+      "description": "Number of requests in RUNNING, WAITING, and SWAPPED state",
+      "fieldConfig": {
+        "defaults": {
+          "color": {
+            "mode": "palette-classic"
+          },
+          "custom": {
+            "axisBorderShow": false,
+            "axisCenteredZero": false,
+            "axisColorMode": "text",
+            "axisLabel": "",
+            "axisPlacement": "auto",
+            "barAlignment": 0,
+            "barWidthFactor": 0.6,
+            "drawStyle": "line",
+            "fillOpacity": 0,
+            "gradientMode": "none",
+            "hideFrom": {
+              "legend": false,
+              "tooltip": false,
+              "viz": false
+            },
+            "insertNulls": false,
+            "lineInterpolation": "linear",
+            "lineWidth": 1,
+            "pointSize": 5,
+            "scaleDistribution": {
+              "type": "linear"
+            },
+            "showPoints": "auto",
+            "showValues": false,
+            "spanNulls": false,
+            "stacking": {
+              "group": "A",
+              "mode": "none"
+            },
+            "thresholdsStyle": {
+              "mode": "off"
+            }
+          },
+          "mappings": [],
+          "thresholds": {
+            "mode": "absolute",
+            "steps": [
+              {
+                "color": "green",
+                "value": 0
+              },
+              {
+                "color": "red",
+                "value": 80
+              }
+            ]
+          },
+          "unit": "none"
+        },
+        "overrides": []
+      },
+      "gridPos": {
+        "h": 8,
+        "w": 12,
+        "x": 12,
+        "y": 8
+      },
+      "id": 3,
+      "options": {
+        "legend": {
+          "calcs": [],
+          "displayMode": "list",
+          "placement": "bottom",
+          "showLegend": true
+        },
+        "tooltip": {
+          "hideZeros": false,
+          "mode": "single",
+          "sort": "none"
+        }
+      },
+      "pluginVersion": "12.2.0",
+      "targets": [
+        {
+          "datasource": {
+            "type": "prometheus",
+            "uid": "${DS_PROMETHEUS}"
+          },
+          "disableTextWrap": false,
+          "editorMode": "builder",
+          "expr": "vllm:num_requests_running{model_name=\"$model_name\"}",
+          "fullMetaSearch": false,
+          "includeNullMetadata": true,
+          "instant": false,
+          "legendFormat": "Num Running",
+          "range": true,
+          "refId": "A",
+          "useBackend": false
+        },
+        {
+          "datasource": {
+            "type": "prometheus",
+            "uid": "${DS_PROMETHEUS}"
+          },
+          "disableTextWrap": false,
+          "editorMode": "builder",
+          "expr": "vllm:num_requests_waiting{model_name=\"$model_name\"}",
+          "fullMetaSearch": false,
+          "hide": false,
+          "includeNullMetadata": true,
+          "instant": false,
+          "legendFormat": "Num Waiting",
+          "range": true,
+          "refId": "C",
+          "useBackend": false
+        }
+      ],
+      "title": "Scheduler State",
+      "type": "timeseries"
+    },
+    {
+      "datasource": {
+        "type": "prometheus",
+        "uid": "${DS_PROMETHEUS}"
+      },
+      "description": "P50, P90, P95, and P99 TTFT latency in seconds.",
+      "fieldConfig": {
+        "defaults": {
+          "color": {
+            "mode": "palette-classic"
+          },
+          "custom": {
+            "axisBorderShow": false,
+            "axisCenteredZero": false,
+            "axisColorMode": "text",
+            "axisLabel": "",
+            "axisPlacement": "auto",
+            "barAlignment": 0,
+            "barWidthFactor": 0.6,
+            "drawStyle": "line",
+            "fillOpacity": 0,
+            "gradientMode": "none",
+            "hideFrom": {
+              "legend": false,
+              "tooltip": false,
+              "viz": false
+            },
+            "insertNulls": false,
+            "lineInterpolation": "linear",
+            "lineWidth": 1,
+            "pointSize": 5,
+            "scaleDistribution": {
+              "type": "linear"
+            },
+            "showPoints": "auto",
+            "showValues": false,
+            "spanNulls": false,
+            "stacking": {
+              "group": "A",
+              "mode": "none"
+            },
+            "thresholdsStyle": {
+              "mode": "off"
+            }
+          },
+          "mappings": [],
+          "thresholds": {
+            "mode": "absolute",
+            "steps": [
+              {
+                "color": "green",
+                "value": 0
+              },
+              {
+                "color": "red",
+                "value": 80
+              }
+            ]
+          },
+          "unit": "s"
+        },
+        "overrides": []
+      },
+      "gridPos": {
+        "h": 8,
+        "w": 12,
+        "x": 0,
+        "y": 16
+      },
+      "id": 5,
+      "interval": "60s",
+      "options": {
+        "legend": {
+          "calcs": [],
+          "displayMode": "list",
+          "placement": "bottom",
+          "showLegend": true
+        },
+        "tooltip": {
+          "hideZeros": false,
+          "mode": "single",
+          "sort": "none"
+        }
+      },
+      "pluginVersion": "12.2.0",
+      "targets": [
+        {
+          "datasource": {
+            "type": "prometheus",
+            "uid": "${DS_PROMETHEUS}"
+          },
+          "disableTextWrap": false,
+          "editorMode": "builder",
+          "expr": "histogram_quantile(0.99, sum by(le) (rate(vllm:time_to_first_token_seconds_bucket{model_name=\"$model_name\"}[$__rate_interval])))",
+          "fullMetaSearch": false,
+          "hide": false,
+          "includeNullMetadata": false,
+          "instant": false,
+          "legendFormat": "P99",
+          "range": true,
+          "refId": "A",
+          "useBackend": false
+        },
+        {
+          "datasource": {
+            "type": "prometheus",
+            "uid": "${DS_PROMETHEUS}"
+          },
+          "disableTextWrap": false,
+          "editorMode": "builder",
+          "expr": "histogram_quantile(0.95, sum by(le) (rate(vllm:time_to_first_token_seconds_bucket{model_name=\"$model_name\"}[$__rate_interval])))",
+          "fullMetaSearch": false,
+          "includeNullMetadata": false,
+          "instant": false,
+          "legendFormat": "P95",
+          "range": true,
+          "refId": "B",
+          "useBackend": false
+        },
+        {
+          "datasource": {
+            "type": "prometheus",
+            "uid": "${DS_PROMETHEUS}"
+          },
+          "disableTextWrap": false,
+          "editorMode": "builder",
+          "expr": "histogram_quantile(0.9, sum by(le) (rate(vllm:time_to_first_token_seconds_bucket{model_name=\"$model_name\"}[$__rate_interval])))",
+          "fullMetaSearch": false,
+          "hide": false,
+          "includeNullMetadata": false,
+          "instant": false,
+          "legendFormat": "P90",
+          "range": true,
+          "refId": "C",
+          "useBackend": false
+        },
+        {
+          "datasource": {
+            "type": "prometheus",
+            "uid": "${DS_PROMETHEUS}"
+          },
+          "disableTextWrap": false,
+          "editorMode": "builder",
+          "expr": "histogram_quantile(0.5, sum by(le) (rate(vllm:time_to_first_token_seconds_bucket{model_name=\"$model_name\"}[$__rate_interval])))",
+          "fullMetaSearch": false,
+          "hide": false,
+          "includeNullMetadata": false,
+          "instant": false,
+          "legendFormat": "P50",
+          "range": true,
+          "refId": "D",
+          "useBackend": false
+        },
+        {
+          "datasource": {
+            "type": "prometheus",
+            "uid": "${DS_PROMETHEUS}"
+          },
+          "editorMode": "code",
+          "expr": "rate(vllm:time_to_first_token_seconds_sum{model_name=\"$model_name\"}[$__rate_interval])\n/\nrate(vllm:time_to_first_token_seconds_count{model_name=\"$model_name\"}[$__rate_interval])",
+          "hide": false,
+          "instant": false,
+          "legendFormat": "Average",
+          "range": true,
+          "refId": "E"
+        }
+      ],
+      "title": "Time To First Token Latency",
+      "type": "timeseries"
+    },
+    {
+      "datasource": {
+        "type": "prometheus",
+        "uid": "${DS_PROMETHEUS}"
+      },
+      "description": "Percentage of used cache blocks by vLLM.",
+      "fieldConfig": {
+        "defaults": {
+          "color": {
+            "mode": "palette-classic"
+          },
+          "custom": {
+            "axisBorderShow": false,
+            "axisCenteredZero": false,
+            "axisColorMode": "text",
+            "axisLabel": "",
+            "axisPlacement": "auto",
+            "barAlignment": 0,
+            "barWidthFactor": 0.6,
+            "drawStyle": "line",
+            "fillOpacity": 0,
+            "gradientMode": "none",
+            "hideFrom": {
+              "legend": false,
+              "tooltip": false,
+              "viz": false
+            },
+            "insertNulls": false,
+            "lineInterpolation": "linear",
+            "lineWidth": 1,
+            "pointSize": 5,
+            "scaleDistribution": {
+              "type": "linear"
+            },
+            "showPoints": "auto",
+            "showValues": false,
+            "spanNulls": false,
+            "stacking": {
+              "group": "A",
+              "mode": "none"
+            },
+            "thresholdsStyle": {
+              "mode": "off"
+            }
+          },
+          "mappings": [],
+          "thresholds": {
+            "mode": "absolute",
+            "steps": [
+              {
+                "color": "green",
+                "value": 0
+              },
+              {
+                "color": "red",
+                "value": 80
+              }
+            ]
+          },
+          "unit": "percentunit"
+        },
+        "overrides": []
+      },
+      "gridPos": {
+        "h": 8,
+        "w": 12,
+        "x": 12,
+        "y": 16
+      },
+      "id": 4,
+      "interval": "60s",
+      "options": {
+        "legend": {
+          "calcs": [],
+          "displayMode": "list",
+          "placement": "bottom",
+          "showLegend": true
+        },
+        "tooltip": {
+          "hideZeros": false,
+          "mode": "single",
+          "sort": "none"
+        }
+      },
+      "pluginVersion": "12.2.0",
+      "targets": [
+        {
+          "datasource": {
+            "type": "prometheus",
+            "uid": "${DS_PROMETHEUS}"
+          },
+          "editorMode": "code",
+          "expr": "vllm:gpu_cache_usage_perc{model_name=\"$model_name\"}",
+          "instant": false,
+          "legendFormat": "GPU Cache Usage",
+          "range": true,
+          "refId": "A"
+        }
+      ],
+      "title": "Cache Utilization",
+      "type": "timeseries"
+    },
+    {
+      "datasource": {
+        "type": "prometheus",
+        "uid": "${DS_PROMETHEUS}"
+      },
+      "description": "Heatmap of request prompt length",
+      "fieldConfig": {
+        "defaults": {
+          "custom": {
+            "hideFrom": {
+              "legend": false,
+              "tooltip": false,
+              "viz": false
+            },
+            "scaleDistribution": {
+              "type": "linear"
+            }
+          }
+        },
+        "overrides": []
+      },
+      "gridPos": {
+        "h": 8,
+        "w": 12,
+        "x": 0,
+        "y": 24
+      },
+      "id": 12,
+      "interval": "60s",
+      "options": {
+        "calculate": false,
+        "cellGap": 1,
+        "cellValues": {
+          "unit": "none"
+        },
+        "color": {
+          "exponent": 0.5,
+          "fill": "dark-orange",
+          "min": 0,
+          "mode": "scheme",
+          "reverse": false,
+          "scale": "exponential",
+          "scheme": "Spectral",
+          "steps": 64
+        },
+        "exemplars": {
+          "color": "rgba(255,0,255,0.7)"
+        },
+        "filterValues": {
+          "le": 1e-9
+        },
+        "legend": {
+          "show": true
+        },
+        "rowsFrame": {
+          "layout": "auto",
+          "value": "Request count"
+        },
+        "tooltip": {
+          "mode": "single",
+          "showColorScale": false,
+          "yHistogram": true
+        },
+        "yAxis": {
+          "axisLabel": "Prompt Length",
+          "axisPlacement": "left",
+          "reverse": false,
+          "unit": "none"
+        }
+      },
+      "pluginVersion": "12.2.0",
+      "targets": [
+        {
+          "datasource": {
+            "type": "prometheus",
+            "uid": "${DS_PROMETHEUS}"
+          },
+          "disableTextWrap": false,
+          "editorMode": "builder",
+          "expr": "sum by(le) (increase(vllm:request_prompt_tokens_bucket{model_name=\"$model_name\"}[$__rate_interval]))",
+          "format": "heatmap",
+          "fullMetaSearch": false,
+          "includeNullMetadata": true,
+          "instant": false,
+          "legendFormat": "{{le}}",
+          "range": true,
+          "refId": "A",
+          "useBackend": false
+        }
+      ],
+      "title": "Request Prompt Length",
+      "type": "heatmap"
+    },
+    {
+      "datasource": {
+        "type": "prometheus",
+        "uid": "${DS_PROMETHEUS}"
+      },
+      "description": "Heatmap of request generation length",
+      "fieldConfig": {
+        "defaults": {
+          "custom": {
+            "hideFrom": {
+              "legend": false,
+              "tooltip": false,
+              "viz": false
+            },
+            "scaleDistribution": {
+              "type": "linear"
+            }
+          }
+        },
+        "overrides": []
+      },
+      "gridPos": {
+        "h": 8,
+        "w": 12,
+        "x": 12,
+        "y": 24
+      },
+      "id": 13,
+      "interval": "60s",
+      "options": {
+        "calculate": false,
+        "cellGap": 1,
+        "cellValues": {
+          "unit": "none"
+        },
+        "color": {
+          "exponent": 0.5,
+          "fill": "dark-orange",
+          "min": 0,
+          "mode": "scheme",
+          "reverse": false,
+          "scale": "exponential",
+          "scheme": "Spectral",
+          "steps": 64
+        },
+        "exemplars": {
+          "color": "rgba(255,0,255,0.7)"
+        },
+        "filterValues": {
+          "le": 1e-9
+        },
+        "legend": {
+          "show": true
+        },
+        "rowsFrame": {
+          "layout": "auto",
+          "value": "Request count"
+        },
+        "tooltip": {
+          "mode": "single",
+          "showColorScale": false,
+          "yHistogram": true
+        },
+        "yAxis": {
+          "axisLabel": "Generation Length",
+          "axisPlacement": "left",
+          "reverse": false,
+          "unit": "none"
+        }
+      },
+      "pluginVersion": "12.2.0",
+      "targets": [
+        {
+          "datasource": {
+            "type": "prometheus",
+            "uid": "${DS_PROMETHEUS}"
+          },
+          "disableTextWrap": false,
+          "editorMode": "builder",
+          "expr": "sum by(le) (increase(vllm:request_generation_tokens_bucket{model_name=\"$model_name\"}[$__rate_interval]))",
+          "format": "heatmap",
+          "fullMetaSearch": false,
+          "includeNullMetadata": true,
+          "instant": false,
+          "legendFormat": "{{le}}",
+          "range": true,
+          "refId": "A",
+          "useBackend": false
+        }
+      ],
+      "title": "Request Generation Length",
+      "type": "heatmap"
+    },
+    {
+      "datasource": {
+        "type": "prometheus",
+        "uid": "${DS_PROMETHEUS}"
+      },
+      "description": "Number of finished requests by their finish reason: either an EOS token was generated or the max sequence length was reached.",
+      "fieldConfig": {
+        "defaults": {
+          "color": {
+            "mode": "palette-classic"
+          },
+          "custom": {
+            "axisBorderShow": false,
+            "axisCenteredZero": false,
+            "axisColorMode": "text",
+            "axisLabel": "",
+            "axisPlacement": "auto",
+            "barAlignment": 0,
+            "barWidthFactor": 0.6,
+            "drawStyle": "line",
+            "fillOpacity": 0,
+            "gradientMode": "none",
+            "hideFrom": {
+              "legend": false,
+              "tooltip": false,
+              "viz": false
+            },
+            "insertNulls": false,
+            "lineInterpolation": "linear",
+            "lineWidth": 1,
+            "pointSize": 5,
+            "scaleDistribution": {
+              "type": "linear"
+            },
+            "showPoints": "auto",
+            "showValues": false,
+            "spanNulls": false,
+            "stacking": {
+              "group": "A",
+              "mode": "none"
+            },
+            "thresholdsStyle": {
+              "mode": "off"
+            }
+          },
+          "mappings": [],
+          "thresholds": {
+            "mode": "absolute",
+            "steps": [
+              {
+                "color": "green",
+                "value": 0
+              },
+              {
+                "color": "red",
+                "value": 80
+              }
+            ]
+          }
+        },
+        "overrides": []
+      },
+      "gridPos": {
+        "h": 8,
+        "w": 12,
+        "x": 0,
+        "y": 32
+      },
+      "id": 11,
+      "interval": "60s",
+      "options": {
+        "legend": {
+          "calcs": [],
+          "displayMode": "list",
+          "placement": "bottom",
+          "showLegend": true
+        },
+        "tooltip": {
+          "hideZeros": false,
+          "mode": "single",
+          "sort": "none"
+        }
+      },
+      "pluginVersion": "12.2.0",
+      "targets": [
+        {
+          "datasource": {
+            "type": "prometheus",
+            "uid": "${DS_PROMETHEUS}"
+          },
+          "disableTextWrap": false,
+          "editorMode": "builder",
+          "expr": "sum by(finished_reason) (increase(vllm:request_success_total{model_name=\"$model_name\"}[$__rate_interval]))",
+          "fullMetaSearch": false,
+          "includeNullMetadata": true,
+          "instant": false,
+          "interval": "",
+          "legendFormat": "__auto",
+          "range": true,
+          "refId": "A",
+          "useBackend": false
+        }
+      ],
+      "title": "Finish Reason",
+      "type": "timeseries"
+    },
+    {
+      "datasource": {
+        "default": false,
+        "type": "prometheus",
+        "uid": "${DS_PROMETHEUS}"
+      },
+      "fieldConfig": {
+        "defaults": {
+          "color": {
+            "mode": "palette-classic"
+          },
+          "custom": {
+            "axisBorderShow": false,
+            "axisCenteredZero": false,
+            "axisColorMode": "text",
+            "axisLabel": "seconds",
+            "axisPlacement": "auto",
+            "barAlignment": 0,
+            "barWidthFactor": 0.6,
+            "drawStyle": "line",
+            "fillOpacity": 0,
+            "gradientMode": "none",
+            "hideFrom": {
+              "legend": false,
+              "tooltip": false,
+              "viz": false
+            },
+            "insertNulls": false,
+            "lineInterpolation": "linear",
+            "lineWidth": 1,
+            "pointSize": 5,
+            "scaleDistribution": {
+              "type": "linear"
+            },
+            "showPoints": "auto",
+            "showValues": false,
+            "spanNulls": false,
+            "stacking": {
+              "group": "A",
+              "mode": "none"
+            },
+            "thresholdsStyle": {
+              "mode": "off"
+            }
+          },
+          "mappings": [],
+          "thresholds": {
+            "mode": "absolute",
+            "steps": [
+              {
+                "color": "green",
+                "value": 0
+              },
+              {
+                "color": "red",
+                "value": 80
+              }
+            ]
+          }
+        },
+        "overrides": []
+      },
+      "gridPos": {
+        "h": 8,
+        "w": 12,
+        "x": 12,
+        "y": 32
+      },
+      "id": 14,
+      "interval": "60s",
+      "options": {
+        "legend": {
+          "calcs": [],
+          "displayMode": "list",
+          "placement": "bottom",
+          "showLegend": true
+        },
+        "tooltip": {
+          "hideZeros": false,
+          "mode": "single",
+          "sort": "none"
+        }
+      },
+      "pluginVersion": "12.2.0",
+      "targets": [
+        {
+          "datasource": {
+            "type": "prometheus",
+            "uid": "${DS_PROMETHEUS}"
+          },
+          "disableTextWrap": false,
+          "editorMode": "code",
+          "expr": "rate(vllm:request_queue_time_seconds_sum{model_name=\"$model_name\"}[$__rate_interval])",
+          "fullMetaSearch": false,
+          "includeNullMetadata": true,
+          "instant": false,
+          "legendFormat": "__auto",
+          "range": true,
+          "refId": "A",
+          "useBackend": false
+        }
+      ],
+      "title": "Queue Time",
+      "type": "timeseries"
+    },
+    {
+      "datasource": {
+        "default": false,
+        "type": "prometheus",
+        "uid": "${DS_PROMETHEUS}"
+      },
+      "fieldConfig": {
+        "defaults": {
+          "color": {
+            "mode": "palette-classic"
+          },
+          "custom": {
+            "axisBorderShow": false,
+            "axisCenteredZero": false,
+            "axisColorMode": "text",
+            "axisLabel": "",
+            "axisPlacement": "auto",
+            "barAlignment": 0,
+            "barWidthFactor": 0.6,
+            "drawStyle": "line",
+            "fillOpacity": 0,
+            "gradientMode": "none",
+            "hideFrom": {
+              "legend": false,
+              "tooltip": false,
+              "viz": false
+            },
+            "insertNulls": false,
+            "lineInterpolation": "linear",
+            "lineWidth": 1,
+            "pointSize": 5,
+            "scaleDistribution": {
+              "type": "linear"
+            },
+            "showPoints": "auto",
+            "showValues": false,
+            "spanNulls": false,
+            "stacking": {
+              "group": "A",
+              "mode": "none"
+            },
+            "thresholdsStyle": {
+              "mode": "off"
+            }
+          },
+          "mappings": [],
+          "thresholds": {
+            "mode": "absolute",
+            "steps": [
+              {
+                "color": "green",
+                "value": 0
+              },
+              {
+                "color": "red",
+                "value": 80
+              }
+            ]
+          }
+        },
+        "overrides": []
+      },
+      "gridPos": {
+        "h": 8,
+        "w": 12,
+        "x": 0,
+        "y": 40
+      },
+      "id": 15,
+      "interval": "60s",
+      "options": {
+        "legend": {
+          "calcs": [],
+          "displayMode": "list",
+          "placement": "bottom",
+          "showLegend": true
+        },
+        "tooltip": {
+          "hideZeros": false,
+          "mode": "single",
+          "sort": "none"
+        }
+      },
+      "pluginVersion": "12.2.0",
+      "targets": [
+        {
+          "datasource": {
+            "type": "prometheus",
+            "uid": "${DS_PROMETHEUS}"
+          },
+          "disableTextWrap": false,
+          "editorMode": "code",
+          "expr": "rate(vllm:request_prefill_time_seconds_sum{model_name=\"$model_name\"}[$__rate_interval])",
+          "fullMetaSearch": false,
+          "includeNullMetadata": true,
+          "instant": false,
+          "legendFormat": "Prefill",
+          "range": true,
+          "refId": "A",
+          "useBackend": false
+        },
+        {
+          "datasource": {
+            "type": "prometheus",
+            "uid": "${DS_PROMETHEUS}"
+          },
+          "editorMode": "code",
+          "expr": "rate(vllm:request_decode_time_seconds_sum{model_name=\"$model_name\"}[$__rate_interval])",
+          "hide": false,
+          "instant": false,
+          "legendFormat": "Decode",
+          "range": true,
+          "refId": "B"
+        }
+      ],
+      "title": "Requests Prefill and Decode Time",
+      "type": "timeseries"
+    },
+    {
+      "datasource": {
+        "default": false,
+        "type": "prometheus",
+        "uid": "${DS_PROMETHEUS}"
+      },
+      "fieldConfig": {
+        "defaults": {
+          "color": {
+            "mode": "palette-classic"
+          },
+          "custom": {
+            "axisBorderShow": false,
+            "axisCenteredZero": false,
+            "axisColorMode": "text",
+            "axisLabel": "",
+            "axisPlacement": "auto",
+            "barAlignment": 0,
+            "barWidthFactor": 0.6,
+            "drawStyle": "line",
+            "fillOpacity": 0,
+            "gradientMode": "none",
+            "hideFrom": {
+              "legend": false,
+              "tooltip": false,
+              "viz": false
+            },
+            "insertNulls": false,
+            "lineInterpolation": "linear",
+            "lineWidth": 1,
+            "pointSize": 5,
+            "scaleDistribution": {
+              "type": "linear"
+            },
+            "showPoints": "auto",
+            "showValues": false,
+            "spanNulls": false,
+            "stacking": {
+              "group": "A",
+              "mode": "none"
+            },
+            "thresholdsStyle": {
+              "mode": "off"
+            }
+          },
+          "mappings": [],
+          "thresholds": {
+            "mode": "absolute",
+            "steps": [
+              {
+                "color": "green",
+                "value": 0
+              },
+              {
+                "color": "red",
+                "value": 80
+              }
+            ]
+          }
+        },
+        "overrides": []
+      },
+      "gridPos": {
+        "h": 8,
+        "w": 12,
+        "x": 12,
+        "y": 40
+      },
+      "id": 16,
+      "interval": "60s",
+      "options": {
+        "legend": {
+          "calcs": [],
+          "displayMode": "list",
+          "placement": "bottom",
+          "showLegend": true
+        },
+        "tooltip": {
+          "hideZeros": false,
+          "mode": "single",
+          "sort": "none"
+        }
+      },
+      "pluginVersion": "12.2.0",
+      "targets": [
+        {
+          "datasource": {
+            "type": "prometheus",
+            "uid": "${DS_PROMETHEUS}"
+          },
+          "disableTextWrap": false,
+          "editorMode": "code",
+          "expr": "rate(vllm:request_max_num_generation_tokens_sum{model_name=\"$model_name\"}[$__rate_interval])",
+          "fullMetaSearch": false,
+          "includeNullMetadata": true,
+          "instant": false,
+          "legendFormat": "Tokens",
+          "range": true,
+          "refId": "A",
+          "useBackend": false
+        }
+      ],
+      "title": "Max Generation Token in Sequence Group",
+      "type": "timeseries"
+    }
+  ],
+  "preload": false,
+  "refresh": "",
+  "schemaVersion": 42,
+  "tags": [],
+  "templating": {
+    "list": [
+      {
+        "current": {
+          "text": "prometheus",
+          "value": "cf0jbbq5bcwsgb"
+        },
+        "includeAll": false,
+        "label": "datasource",
+        "name": "DS_PROMETHEUS",
+        "options": [],
+        "query": "prometheus",
+        "refresh": 1,
+        "regex": "",
+        "type": "datasource"
+      },
+      {
+        "current": {
+          "text": "meta-llama/Llama-3.2-3B-Instruct",
+          "value": "meta-llama/Llama-3.2-3B-Instruct"
+        },
+        "datasource": {
+          "type": "prometheus",
+          "uid": "${DS_PROMETHEUS}"
+        },
+        "definition": "label_values(model_name)",
+        "includeAll": false,
+        "label": "model_name",
+        "name": "model_name",
+        "options": [],
+        "query": {
+          "query": "label_values(model_name)",
+          "refId": "StandardVariableQuery"
+        },
+        "refresh": 1,
+        "regex": "",
+        "type": "query"
+      }
+    ]
+  },
+  "time": {
+    "from": "now-15m",
+    "to": "now"
+  },
+  "timepicker": {},
+  "timezone": "",
+  "title": "vLLM",
+  "uid": "b281712d-8bff-41ef-9f3f-71ad43c05e9b",
+  "version": 9
+}
\ No newline at end of file
diff --git a/ansible/playbooks/saas/roles/grafana/files/provisioning/plugins/grafana-piechart-panel.yaml b/ansible/playbooks/saas/roles/grafana/files/provisioning/plugins/grafana-piechart-panel.yaml
deleted file mode 100644
index 10ee02f7..00000000
--- a/ansible/playbooks/saas/roles/grafana/files/provisioning/plugins/grafana-piechart-panel.yaml
+++ /dev/null
@@ -1,15 +0,0 @@
-# apiVersion: 1
-
-# apps:
-#   - type: grafana-piechart-panel
-#     org_id: 1
-#     # org_name: Main Org.
-#     disabled: false
-#     # # <map> fields that will be converted to json and stored in jsonData. Custom per app.
-#     # jsonData:
-#     #   # key/value pairs of string to object
-#     #   key: value
-#     # # <map> fields that will be converted to json, encrypted and stored in secureJsonData. Custom per app.
-#     # secureJsonData:
-#     #   # key/value pairs of string to string
-#     #   key: value
diff --git a/ansible/playbooks/saas/roles/grafana/tasks/destroy.yml b/ansible/playbooks/saas/roles/grafana/tasks/destroy.yml
index ce77a12b..e00b7de1 100644
--- a/ansible/playbooks/saas/roles/grafana/tasks/destroy.yml
+++ b/ansible/playbooks/saas/roles/grafana/tasks/destroy.yml
@@ -8,3 +8,4 @@
   ansible.builtin.file:
     path: "{{ software_path }}"
     state: absent
+  delegate_to: "{{ software.instance }}"
\ No newline at end of file
diff --git a/ansible/playbooks/saas/roles/grafana/tasks/main.yml b/ansible/playbooks/saas/roles/grafana/tasks/main.yml
index 3211d5c1..32f26abd 100644
--- a/ansible/playbooks/saas/roles/grafana/tasks/main.yml
+++ b/ansible/playbooks/saas/roles/grafana/tasks/main.yml
@@ -1,4 +1,15 @@
 ---
+- name: Create default directory
+  ansible.builtin.file:
+    path: "{{ item }}"
+    state: directory
+    owner: root
+    group: root
+    mode: "0755"
+  loop:
+    - "{{ software_path }}"
+  delegate_to: "{{ software.instance }}"
+
 - name: Copy Grafana content files
   ansible.builtin.copy:
     src: "{{ item }}"
@@ -9,16 +20,21 @@
   loop:
     - dashboards
     - provisioning
+  delegate_to: "{{ software.instance }}"
 
 - name: Copy Grafana content templates
   ansible.builtin.template:
-    src: "provisioning/datasources/{{ item }}.j2"
-    dest: "{{ software_path }}/provisioning/datasources/{{ item }}"
+    src: "provisioning/{{ item.path }}/{{ item.file }}.j2"
+    dest: "{{ software_path }}/provisioning/{{ item.path }}/{{ item.file }}"
     owner: root
     group: root
     mode: '0644'
   loop:
-    - prometheus.yaml
+    - path: datasources
+      file: prometheus.yaml
+    - path: plugins
+      file: llm.yaml
+  delegate_to: "{{ software.instance }}"
 
 - name: Copy nomad job
   ansible.builtin.template:
diff --git a/ansible/playbooks/saas/roles/grafana/templates/nomad.hcl b/ansible/playbooks/saas/roles/grafana/templates/nomad.hcl
index a17e4205..2b68803f 100644
--- a/ansible/playbooks/saas/roles/grafana/templates/nomad.hcl
+++ b/ansible/playbooks/saas/roles/grafana/templates/nomad.hcl
@@ -3,7 +3,7 @@ job "{{ domain }}" {
   datacenters = ["{{ fact_instance.datacenter }}"]
   type = "service"
 
-{% if software.constraints.location %}
+{% if software.constraints is defined and software.constraints.location is defined %}
   constraint {
     attribute    = "${meta.location}"
     set_contains = "{{ software.constraints.location }}"
@@ -41,7 +41,7 @@ job "{{ domain }}" {
       env {
         GF_LOG_MODE = "console"
         GF_SERVER_HTTP_PORT = "3000"
-        GF_INSTALL_PLUGINS = "grafana-piechart-panel"
+        GF_INSTALL_PLUGINS = "grafana-piechart-panel,grafana-llm-app"
         GF_SECURITY_ADMIN_USER = "{{ lookup('simple-stack-ui', type='secret', key=domain, subkey='user', missing='create', nosymbols=true, length=8) }}"
         GF_SECURITY_ADMIN_PASSWORD = "{{ lookup('simple-stack-ui', type='secret', key=domain, subkey='passwd', missing='create', length=12) }}"
         DS_PROMETHEUS = "prometheus"
diff --git a/ansible/playbooks/saas/roles/grafana/templates/provisioning/datasources/prometheus.yaml.j2 b/ansible/playbooks/saas/roles/grafana/templates/provisioning/datasources/prometheus.yaml.j2
index 81abd7e0..539aa307 100644
--- a/ansible/playbooks/saas/roles/grafana/templates/provisioning/datasources/prometheus.yaml.j2
+++ b/ansible/playbooks/saas/roles/grafana/templates/provisioning/datasources/prometheus.yaml.j2
@@ -1,5 +1,6 @@
 apiVersion: 1
 
+{% if prometheus_remote_write is defined %}
 datasources:
   - name: Mimir
     uid: prometheus
@@ -47,3 +48,6 @@ datasources:
       basicAuthPassword: "{{ loki_remote_write.password }}"
     isDefault: false
 {% endif %}
+{% else %}
+datasources: []
+{% endif %}
\ No newline at end of file
diff --git a/ansible/playbooks/saas/roles/grafana/templates/provisioning/plugins/llm.yaml.j2 b/ansible/playbooks/saas/roles/grafana/templates/provisioning/plugins/llm.yaml.j2
new file mode 100644
index 00000000..2d5cdd9e
--- /dev/null
+++ b/ansible/playbooks/saas/roles/grafana/templates/provisioning/plugins/llm.yaml.j2
@@ -0,0 +1,4 @@
+apiVersion: 1
+
+apps:
+{{ (lookup('simple-stack-ui', type='secret', key=domain, subkey='plugins', missing='error') | from_json) | to_nice_yaml }}
\ No newline at end of file
diff --git a/ansible/playbooks/saas/roles/homeassistant/templates/nomad.hcl b/ansible/playbooks/saas/roles/homeassistant/templates/nomad.hcl
index ec5ba3a6..68e101f1 100644
--- a/ansible/playbooks/saas/roles/homeassistant/templates/nomad.hcl
+++ b/ansible/playbooks/saas/roles/homeassistant/templates/nomad.hcl
@@ -3,7 +3,7 @@ job "{{ domain }}" {
   datacenters = ["{{ fact_instance.datacenter }}"]
   type = "service"
 
-{% if software.constraints.location %}
+{% if software.constraints is defined and software.constraints.location is defined %}
   constraint {
     attribute    = "${meta.location}"
     set_contains = "{{ software.constraints.location }}"
diff --git a/ansible/playbooks/saas/roles/kresus/templates/nomad.hcl b/ansible/playbooks/saas/roles/kresus/templates/nomad.hcl
index 75da59ff..1e8d7e7d 100644
--- a/ansible/playbooks/saas/roles/kresus/templates/nomad.hcl
+++ b/ansible/playbooks/saas/roles/kresus/templates/nomad.hcl
@@ -3,7 +3,7 @@ job "{{ domain }}" {
   datacenters = ["{{ fact_instance.datacenter }}"]
   type = "service"
 
-{% if software.constraints.location %}
+{% if software.constraints is defined and software.constraints.location is defined %}
   constraint {
     attribute    = "${meta.location}"
     set_contains = "{{ software.constraints.location }}"
diff --git a/ansible/playbooks/saas/roles/loki/tasks/main.yml b/ansible/playbooks/saas/roles/loki/tasks/main.yml
index e10eec4a..10a2e33a 100644
--- a/ansible/playbooks/saas/roles/loki/tasks/main.yml
+++ b/ansible/playbooks/saas/roles/loki/tasks/main.yml
@@ -8,15 +8,7 @@
     mode: '0755'
   loop:
     - "{{ software_path }}/var/lib/loki"
-    - "{{ software_path }}/etc/loki"
-
-- name: Copy config file
-  ansible.builtin.template:
-    src: config.yaml.j2
-    dest: "{{ software_path }}/etc/loki/local-config.yaml"
-    owner: 10001
-    group: 10001
-    mode: '0644'
+  delegate_to: "{{ software.instance }}"
 
 - name: Copy nomad job
   ansible.builtin.template:
diff --git a/ansible/playbooks/saas/roles/loki/templates/nomad.hcl b/ansible/playbooks/saas/roles/loki/templates/nomad.hcl
index ed695b61..514509d3 100644
--- a/ansible/playbooks/saas/roles/loki/templates/nomad.hcl
+++ b/ansible/playbooks/saas/roles/loki/templates/nomad.hcl
@@ -3,7 +3,7 @@ job "{{ domain }}" {
   datacenters = ["{{ fact_instance.datacenter }}"]
   type = "service"
 
-{% if software.constraints.location %}
+{% if software.constraints is defined and software.constraints.location is defined %}
   constraint {
     attribute    = "${meta.location}"
     set_contains = "{{ software.constraints.location }}"
@@ -40,12 +40,23 @@ job "{{ domain }}" {
       config {
         image = "grafana/loki:{{ softwares.loki.version }}"
         volumes = [
-          "{{ software_path }}/var/lib/loki:/var/lib/loki:rw",
-          "{{ software_path }}/etc/loki:/etc/loki:ro"
+          "{{ software_path }}/var/lib/loki:/var/lib/loki:rw"
+        ]
+        args = [
+          "-config.file",
+          "/local/config.yaml"
         ]
         ports = ["loki"]
       }
 
+      template {
+        change_mode = "restart"
+        destination = "local/config.yaml"
+        data = <<EOH
+{{ lookup('ansible.builtin.template', 'templates/config.yaml.j2') }}
+  EOH
+      }
+
       resources {
         cpu    = {{ size[software.size].cpu }}
         memory = {{ size[software.size].memory }}
diff --git a/ansible/playbooks/saas/roles/mariadb/templates/nomad.hcl b/ansible/playbooks/saas/roles/mariadb/templates/nomad.hcl
index c21e91aa..ffd8ebe5 100644
--- a/ansible/playbooks/saas/roles/mariadb/templates/nomad.hcl
+++ b/ansible/playbooks/saas/roles/mariadb/templates/nomad.hcl
@@ -3,7 +3,7 @@ job "{{ domain }}" {
   datacenters = ["{{ fact_instance.datacenter }}"]
   type = "service"
 
-{% if software.constraints.location %}
+{% if software.constraints is defined and software.constraints.location is defined %}
   constraint {
     attribute    = "${meta.location}"
     set_contains = "{{ software.constraints.location }}"
diff --git a/ansible/playbooks/saas/roles/mimir/tasks/main.yml b/ansible/playbooks/saas/roles/mimir/tasks/main.yml
index 342e53e7..f4de5a9f 100644
--- a/ansible/playbooks/saas/roles/mimir/tasks/main.yml
+++ b/ansible/playbooks/saas/roles/mimir/tasks/main.yml
@@ -1,39 +1,4 @@
 ---
-- name: Create data directories
-  ansible.builtin.file:
-    path: "{{ item.path }}"
-    state: directory
-    owner: "{{ item.owner | default('root') }}"
-    group: "{{ item.group | default('root') }}"
-    mode: '0755'
-  loop:
-    - path: /data/{{ domain }}/config
-    - path: /data/{{ domain }}/prometheus
-      owner: nobody
-      group: nogroup
-
-- name: Copy services configurations files
-  ansible.builtin.copy:
-    src: "{{ item }}"
-    dest: "/data/{{ domain }}/config/{{ item }}"
-    owner: root
-    group: root
-    mode: '0644'
-  loop:
-    - prometheus.yaml
-    - rules.yaml
-
-- name: Template services configurations files
-  ansible.builtin.template:
-    src: "{{ item }}.j2"
-    dest: "/data/{{ domain }}/config/{{ item }}"
-    owner: root
-    group: root
-    mode: '0644'
-  loop:
-    - mimir.yaml
-    - alertmanager-fallback-config.yaml
-
 - name: Copy nomad job
   ansible.builtin.template:
     src: nomad.hcl
diff --git a/ansible/playbooks/saas/roles/mimir/templates/mimir.yaml.j2 b/ansible/playbooks/saas/roles/mimir/templates/mimir.yaml.j2
index 8c13f86d..2083a297 100644
--- a/ansible/playbooks/saas/roles/mimir/templates/mimir.yaml.j2
+++ b/ansible/playbooks/saas/roles/mimir/templates/mimir.yaml.j2
@@ -4,13 +4,20 @@ target: all,alertmanager,overrides-exporter
 limits:
   compactor_blocks_retention_period: 90d
 
+common:
+  storage:
+    backend: s3
+    s3:
+      endpoint: minio-{{ service_name }}.default.service.nomad:9000
+      access_key_id: "{{ lookup('simple-stack-ui', type='secret', key=domain, subkey='access_key_id', missing='error') }}"
+      secret_access_key: "{{ lookup('simple-stack-ui', type='secret', key=domain, subkey='secret_access_key', missing='error') }}"
+      insecure: true
+
+memberlist:
+  advertise_addr: {% raw %}{{env "attr.unique.network.ip-address"}}{% endraw %}
+
 blocks_storage:
-  backend: s3
   s3:
-    endpoint: 172.17.0.1:9000
-    access_key_id: "{{ lookup('simple-stack-ui', type='secret', key=domain, subkey='access_key_id', missing='create', length=12) }}"
-    secret_access_key: "{{ lookup('simple-stack-ui', type='secret', key=domain, subkey='secret_access_key', missing='create', length=12) }}"
-    insecure: true
     bucket_name: mimir-blocks
 
 ruler:
@@ -18,31 +25,46 @@ ruler:
   ring:
     heartbeat_period: 2s
     heartbeat_timeout: 10s
+    instance_addr: {% raw %}{{env "attr.unique.network.ip-address"}}{% endraw %}
+
 
 ruler_storage:
-  backend: s3
   s3:
-    endpoint: 172.17.0.1:9000
-    access_key_id: "{{ lookup('simple-stack-ui', type='secret', key=domain, subkey='access_key_id', missing='error') }}"
-    secret_access_key: "{{ lookup('simple-stack-ui', type='secret', key=domain, subkey='secret_access_key', missing='error') }}"
-    insecure: true
     bucket_name: ruler
 
 alertmanager_storage:
-  backend: s3
   s3:
-    endpoint: 172.17.0.1:9000
-    access_key_id: "{{ lookup('simple-stack-ui', type='secret', key=domain, subkey='access_key_id', missing='error') }}"
-    secret_access_key: "{{ lookup('simple-stack-ui', type='secret', key=domain, subkey='secret_access_key', missing='error') }}"
-    insecure: true
     bucket_name: mimir
 
 alertmanager:
-  fallback_config_file: /config/alertmanager-fallback-config.yaml
+  fallback_config_file: /local/alertmanager-fallback-config.yaml
   external_url: http://127.0.0.1:8080/alertmanager
+  sharding_ring:
+    instance_addr: {% raw %}{{env "attr.unique.network.ip-address"}}{% endraw %}
+
+
+compactor:
+  sharding_ring:
+    instance_addr: {% raw %}{{env "attr.unique.network.ip-address"}}{% endraw %}
+
+
+distributor: 
+  ring:
+    instance_addr: {% raw %}{{env "attr.unique.network.ip-address"}}{% endraw %}
+
 
 server:
-  log_level: info
+  log_level: warn
+
+store_gateway: 
+  sharding_ring:
+    instance_addr: {% raw %}{{env "attr.unique.network.ip-address"}}{% endraw %}
+
+
+ingester: 
+  ring:
+    instance_addr: {% raw %}{{env "attr.unique.network.ip-address"}}{% endraw %}
+
 
 usage_stats:
   enabled: false
diff --git a/ansible/playbooks/saas/roles/mimir/templates/nomad.hcl b/ansible/playbooks/saas/roles/mimir/templates/nomad.hcl
index 66c84ec5..cfd71437 100644
--- a/ansible/playbooks/saas/roles/mimir/templates/nomad.hcl
+++ b/ansible/playbooks/saas/roles/mimir/templates/nomad.hcl
@@ -3,21 +3,30 @@ job "{{ domain }}" {
   datacenters = ["{{ fact_instance.datacenter }}"]
   type        = "service"
 
-{% if software.constraints.location %}
+{% if software.constraints is defined and software.constraints.location is defined %}
   constraint {
     attribute    = "${meta.location}"
     set_contains = "{{ software.constraints.location }}"
   }
 {% endif %}
 
+{% if software.constraints is defined and software.constraints.instance is defined %}
   constraint {
     attribute    = "${meta.instance}"
-    set_contains = "{{ software.instance }}"
+    set_contains = "{{ software.constraints.instance }}"
   }
+{% endif %}
 
   group "{{ domain }}-minio" {
     count = 1
 
+{% if software.constraints is defined and software.constraints.minio_instance is defined %}
+    constraint {
+      attribute    = "${meta.instance}"
+      set_contains = "{{ software.constraints.minio_instance }}"
+    }
+{% endif %}
+
     network {
       port "minio" {
         to = 9000
@@ -36,8 +45,8 @@ job "{{ domain }}" {
       driver = "docker"
 
       env {
-        MINIO_ROOT_USER = "{{ lookup('simple-stack-ui', type='secret', key=domain, subkey='access_key_id', missing='error') }}"
-        MINIO_ROOT_PASSWORD = "{{ lookup('simple-stack-ui', type='secret', key=domain, subkey='secret_access_key', missing='error') }}"
+        MINIO_ROOT_USER = "{{ lookup('simple-stack-ui', type='secret', key=domain, subkey='access_key_id', missing='create', length=12) }}"
+        MINIO_ROOT_PASSWORD = "{{ lookup('simple-stack-ui', type='secret', key=domain, subkey='secret_access_key', missing='create', length=12) }}"
       }
 
       config {
@@ -61,15 +70,25 @@ job "{{ domain }}" {
   group "{{ domain }}-mimir" {
     count = 3
 
+{% if software.constraints is defined and software.constraints.distinct_hosts is defined %}
+    constraint {
+      operator  = "distinct_hosts"
+      value     = "{{ software.constraints.distinct_hosts | lower }}"
+    }
+{% endif %}
+
     network {
       port "mimir_7946" {
         to = 7946
+        static = 7946
       }
       port "mimir_9095" {
         to = 9095
+        static = 9095
       }
       port "mimir_8080" {
         to = 8080
+        static = 8080
       }
     }
 
@@ -98,19 +117,55 @@ job "{{ domain }}" {
 
       driver = "docker"
 
+      env {
+        IP = "${attr.unique.network.ip-address}"
+      }
+
       config {
         image = "grafana/mimir:{{ softwares.mimir.version }}"
+        network_mode = "host"
         volumes = [
-          "{{ software_path }}/config:/config",
           "{{ software_path }}/mimir-${NOMAD_ALLOC_INDEX}:/data",
         ]
         args = [
-          "-config.file=/config/mimir.yaml",
+          "-config.file=/local/mimir.yaml",
           "-memberlist.join=dnssrv+mimir.default.service.nomad"
         ]
         ports = ["mimir_7946", "mimir_9095", "mimir_8080"]
       }
 
+      template {
+        change_mode = "restart"
+        destination = "local/mimir.yaml"
+        data = <<EOH
+{{ lookup('ansible.builtin.template', 'templates/mimir.yaml.j2') }}
+  EOH
+      }
+
+      template {
+        change_mode = "restart"
+        destination = "local/rules.yaml"
+        data = <<EOH
+{{ lookup('ansible.builtin.template', 'files/rules.yaml') }}
+  EOH
+      }
+
+      template {
+        change_mode = "restart"
+        destination = "local/alertmanager-fallback-config.yaml"
+        data = <<EOH
+{{ lookup('ansible.builtin.template', 'templates/alertmanager-fallback-config.yaml.j2') }}
+  EOH
+      }
+
+      template {
+        change_mode = "restart"
+        destination = "local/prometheus.yaml"
+        data = <<EOH
+{{ lookup('ansible.builtin.template', 'templates/prometheus.yaml') }}
+  EOH
+      }
+
       resources {
         cpu    = {{ size[software.size].cpu }}
         memory = {{ size[software.size].memory | int }}
diff --git a/ansible/playbooks/saas/roles/mimir/files/prometheus.yaml b/ansible/playbooks/saas/roles/mimir/templates/prometheus.yaml
similarity index 93%
rename from ansible/playbooks/saas/roles/mimir/files/prometheus.yaml
rename to ansible/playbooks/saas/roles/mimir/templates/prometheus.yaml
index eb58c0c4..8e12aaec 100644
--- a/ansible/playbooks/saas/roles/mimir/files/prometheus.yaml
+++ b/ansible/playbooks/saas/roles/mimir/templates/prometheus.yaml
@@ -5,7 +5,7 @@ global:
     namespace: demo
 
 rule_files:
-  - /etc/prometheus/rules.yaml
+  - /local/rules.yaml
 
 scrape_configs:
   - job_name: demo/mimir
diff --git a/ansible/playbooks/saas/roles/minio/tasks/main.yml b/ansible/playbooks/saas/roles/minio/tasks/main.yml
index 80f982f7..f61aa1fa 100644
--- a/ansible/playbooks/saas/roles/minio/tasks/main.yml
+++ b/ansible/playbooks/saas/roles/minio/tasks/main.yml
@@ -1,15 +1,4 @@
 ---
-- name: Create default directory
-  ansible.builtin.file:
-    path: "{{ item.path }}"
-    state: directory
-    owner: "{{ item.owner | default('root') }}"
-    group: "{{ item.group | default('root') }}"
-    mode: '0755'
-  loop:
-    - path: "{{ software_path }}/data"
-    - path: "{{ software_path }}/var/backup"
-
 - name: Copy nomad job to destination
   ansible.builtin.template:
     src: nomad.hcl
diff --git a/ansible/playbooks/saas/roles/minio/templates/nomad.hcl b/ansible/playbooks/saas/roles/minio/templates/nomad.hcl
index db983a14..e9a92136 100644
--- a/ansible/playbooks/saas/roles/minio/templates/nomad.hcl
+++ b/ansible/playbooks/saas/roles/minio/templates/nomad.hcl
@@ -3,7 +3,7 @@ job "{{ domain }}" {
   datacenters = ["{{ fact_instance.datacenter }}"]
   type = "service"
 
-{% if software.constraints.location %}
+{% if software.constraints is defined and software.constraints.location is defined %}
   constraint {
     attribute    = "${meta.location}"
     set_contains = "{{ software.constraints.location }}"
@@ -21,6 +21,9 @@ job "{{ domain }}" {
     network {
       port "minio" {
         to = 9000
+{% if software.static_port is defined %}
+        static = {{ software.static_port }}
+{% endif %}
       }
     }
 
@@ -53,7 +56,7 @@ job "{{ domain }}" {
       }
 
       config {
-        image = "{{ docker_private_registry.url }}/minio:{{ softwares.minio.version }}"
+        image = "{{ docker_private_registry.url_project | default(docker_private_registry.url) }}/minio:{{ softwares.minio.version }}"
         volumes = [
           "{{ software_path }}/data:/data:rw",
           "{{ software_path }}/var/backup:/var/backup:rw"
diff --git a/ansible/playbooks/saas/roles/mosquitto/templates/nomad.hcl b/ansible/playbooks/saas/roles/mosquitto/templates/nomad.hcl
index 600ce577..44264b10 100644
--- a/ansible/playbooks/saas/roles/mosquitto/templates/nomad.hcl
+++ b/ansible/playbooks/saas/roles/mosquitto/templates/nomad.hcl
@@ -3,7 +3,7 @@ job "{{ domain }}" {
   datacenters = ["{{ fact_instance.datacenter }}"]
   type = "service"
 
-{% if software.constraints.location %}
+{% if software.constraints is defined and software.constraints.location is defined %}
   constraint {
     attribute    = "${meta.location}"
     set_contains = "{{ software.constraints.location }}"
diff --git a/ansible/playbooks/saas/roles/nextcloud/templates/nomad.hcl b/ansible/playbooks/saas/roles/nextcloud/templates/nomad.hcl
index 5fd428f3..4e8bf378 100644
--- a/ansible/playbooks/saas/roles/nextcloud/templates/nomad.hcl
+++ b/ansible/playbooks/saas/roles/nextcloud/templates/nomad.hcl
@@ -3,7 +3,7 @@ job "{{ domain }}" {
   datacenters = ["{{ fact_instance.datacenter }}"]
   type = "service"
 
-{% if software.constraints.location %}
+{% if software.constraints is defined and software.constraints.location is defined %}
   constraint {
     attribute    = "${meta.location}"
     set_contains = "{{ software.constraints.location }}"
diff --git a/ansible/playbooks/saas/roles/nginx/templates/nomad.hcl b/ansible/playbooks/saas/roles/nginx/templates/nomad.hcl
index 2e3166fb..340167f2 100644
--- a/ansible/playbooks/saas/roles/nginx/templates/nomad.hcl
+++ b/ansible/playbooks/saas/roles/nginx/templates/nomad.hcl
@@ -3,7 +3,7 @@ job "{{ domain }}" {
   datacenters = ["{{ fact_instance.datacenter }}"]
   type        = "service"
 
-{% if software.constraints.location %}
+{% if software.constraints is defined and software.constraints.location is defined %}
   constraint {
     attribute    = "${meta.location}"
     set_contains = "{{ software.constraints.location }}"
diff --git a/ansible/playbooks/saas/roles/nomad/templates/backup.hcl.j2 b/ansible/playbooks/saas/roles/nomad/templates/backup.hcl.j2
index d348356c..1c0e4b1a 100644
--- a/ansible/playbooks/saas/roles/nomad/templates/backup.hcl.j2
+++ b/ansible/playbooks/saas/roles/nomad/templates/backup.hcl.j2
@@ -3,7 +3,7 @@ job "{{ nomad_job_name }}" {
   datacenters = ["{{ fact_instance.datacenter }}"]
   type = "batch"
 
-{% if software.constraints.location %}
+{% if software.constraints is defined and software.constraints.location is defined %}
   constraint {
     attribute    = "${meta.location}"
     set_contains = "{{ software.constraints.location }}"
diff --git a/ansible/playbooks/saas/roles/nomad/templates/restore.hcl.j2 b/ansible/playbooks/saas/roles/nomad/templates/restore.hcl.j2
index fb7cae8b..00419a79 100644
--- a/ansible/playbooks/saas/roles/nomad/templates/restore.hcl.j2
+++ b/ansible/playbooks/saas/roles/nomad/templates/restore.hcl.j2
@@ -3,7 +3,7 @@ job "{{ nomad_job_name }}" {
   datacenters = ["{{ fact_instance.datacenter }}"]
   type = "batch"
 
-{% if software.constraints.location %}
+{% if software.constraints is defined and software.constraints.location is defined %}
   constraint {
     attribute    = "${meta.location}"
     set_contains = "{{ software.constraints.location }}"
diff --git a/ansible/playbooks/saas/roles/open-webui/templates/nomad.hcl b/ansible/playbooks/saas/roles/open-webui/templates/nomad.hcl
index bc1e1a92..39a276cf 100644
--- a/ansible/playbooks/saas/roles/open-webui/templates/nomad.hcl
+++ b/ansible/playbooks/saas/roles/open-webui/templates/nomad.hcl
@@ -3,7 +3,7 @@ job "{{ domain }}" {
   datacenters = ["{{ fact_instance.datacenter }}"]
   type = "service"
 
-{% if software.constraints.location %}
+{% if software.constraints is defined and software.constraints.location is defined %}
   constraint {
     attribute    = "${meta.location}"
     set_contains = "{{ software.constraints.location }}"
diff --git a/ansible/playbooks/saas/roles/postgresql/templates/nomad.hcl b/ansible/playbooks/saas/roles/postgresql/templates/nomad.hcl
index b751328a..8bce6575 100644
--- a/ansible/playbooks/saas/roles/postgresql/templates/nomad.hcl
+++ b/ansible/playbooks/saas/roles/postgresql/templates/nomad.hcl
@@ -3,7 +3,7 @@ job "{{ domain }}" {
   datacenters = ["{{ fact_instance.datacenter }}"]
   type        = "service"
 
-{% if software.constraints.location %}
+{% if software.constraints is defined and software.constraints.location is defined %}
   constraint {
     attribute    = "${meta.location}"
     set_contains = "{{ software.constraints.location }}"
diff --git a/ansible/playbooks/saas/roles/registry/templates/nomad.hcl b/ansible/playbooks/saas/roles/registry/templates/nomad.hcl
index cd2ec90f..be570e06 100644
--- a/ansible/playbooks/saas/roles/registry/templates/nomad.hcl
+++ b/ansible/playbooks/saas/roles/registry/templates/nomad.hcl
@@ -7,7 +7,7 @@ job "{{ domain }}" {
   datacenters = ["{{ fact_instance.datacenter }}"]
   type = "service"
 
-{% if software.constraints.location %}
+{% if software.constraints is defined and software.constraints.location is defined %}
   constraint {
     attribute    = "${meta.location}"
     set_contains = "{{ software.constraints.location }}"
@@ -59,6 +59,8 @@ job "{{ domain }}" {
         REGISTRY_AUTH = "htpasswd"
         REGISTRY_AUTH_HTPASSWD_REALM = "Registry Realm"
         REGISTRY_AUTH_HTPASSWD_PATH = "/data/htpasswd"
+        REGISTRY_LOG_LEVEL = "info"
+        OTEL_TRACES_EXPORTER = "none"
         #There is a warning about the *HTTP secret*. It's only important if you have multiples registries behind a load balancer.
         REGISTRY_HTTP_SECRET = "Aumuu5ie-ieX7uwee-Aixah4ee"
       }
diff --git a/ansible/playbooks/saas/roles/rocketchat/templates/nomad.hcl b/ansible/playbooks/saas/roles/rocketchat/templates/nomad.hcl
index 2cc28193..95c78511 100644
--- a/ansible/playbooks/saas/roles/rocketchat/templates/nomad.hcl
+++ b/ansible/playbooks/saas/roles/rocketchat/templates/nomad.hcl
@@ -3,7 +3,7 @@ job "{{ domain }}" {
   datacenters = ["{{ fact_instance.datacenter }}"]
   type = "service"
 
-{% if software.constraints.location %}
+{% if software.constraints is defined and software.constraints.location is defined %}
   constraint {
     attribute    = "${meta.location}"
     set_contains = "{{ software.constraints.location }}"
@@ -114,6 +114,8 @@ EOH
     task "{{ domain }}-rocketchat" {
       driver = "docker"
 
+      user = "42420:42420"
+
       env {
         ROOT_URL = "https://{{ domain }}"
         MONGO_URL = "mongodb://${NOMAD_HOST_ADDR_mongodb}/parties?replicaSet=rs0&directConnection=true"
diff --git a/ansible/playbooks/saas/roles/traefik/templates/nomad.hcl b/ansible/playbooks/saas/roles/traefik/templates/nomad.hcl
index b8065a6c..cf901f86 100644
--- a/ansible/playbooks/saas/roles/traefik/templates/nomad.hcl
+++ b/ansible/playbooks/saas/roles/traefik/templates/nomad.hcl
@@ -3,7 +3,7 @@ job "{{ domain }}" {
   datacenters = ["{{ fact_instance.datacenter }}"]
   type = "service"
 
-{% if software.constraints.location %}
+{% if software.constraints is defined and software.constraints.location is defined %}
   constraint {
     attribute    = "${meta.location}"
     set_contains = "{{ software.constraints.location }}"
diff --git a/ansible/playbooks/saas/roles/traefik/templates/traefik.toml b/ansible/playbooks/saas/roles/traefik/templates/traefik.toml
index 031c8bd1..a6bfce55 100644
--- a/ansible/playbooks/saas/roles/traefik/templates/traefik.toml
+++ b/ansible/playbooks/saas/roles/traefik/templates/traefik.toml
@@ -14,6 +14,8 @@
         [entryPoints.http.http.redirections.entrypoint]
           to = "https"
           scheme = "https"
+  [entryPoints.metrics]
+    address = ":8081"
   [entryPoints.https]
     address = ":443"
 
@@ -25,7 +27,7 @@
     buckets = [0.1,0.3,1.2,5.0]
     addEntryPointsLabels = true
     addServicesLabels = true
-    entryPoint = "https"
+    entryPoint = "metrics"
 
 [log]
   level = "WARN"
@@ -47,7 +49,7 @@
     prefix = "traefik"
     exposedByDefault = false
   [providers.nomad.endpoint]
-    address = "https://172.17.0.1:4646"
+    address = "https://{{ hostvars[nomad_primary_master_node]['ansible_ens3']['ipv4']['address'] | default('127.0.0.1') }}:4646"
     token = "{{ lookup('simple-stack-ui', type='secret', key=inventory_hostname, subkey='nomad_traefik_token', missing='error') }}"
   [providers.nomad.endpoint.tls]
     ca = "/etc/ssl/simplestack/simplestack-ca.pem"
diff --git a/ansible/playbooks/saas/roles/valkey/tasks/main.yml b/ansible/playbooks/saas/roles/valkey/tasks/main.yml
index a3262df1..6bb83999 100644
--- a/ansible/playbooks/saas/roles/valkey/tasks/main.yml
+++ b/ansible/playbooks/saas/roles/valkey/tasks/main.yml
@@ -1,22 +1,4 @@
 ---
-- name: Create default directories
-  ansible.builtin.file:
-    path: "{{ item }}"
-    state: directory
-    owner: root
-    group: root
-    mode: '0755'
-  loop:
-    - "{{ software_path }}/data"
-    - "{{ software_path }}/etc/valkey"
-    - "{{ software_path }}/run/valkey"
-
-- name: Copy config file
-  ansible.builtin.copy:
-    src: valkey.conf
-    dest: "{{ software_path }}/etc/valkey/valkey.conf"
-    mode: '0644'
-
 - name: Copy nomad job to destination
   ansible.builtin.template:
     src: nomad.hcl
diff --git a/ansible/playbooks/saas/roles/valkey/templates/nomad.hcl b/ansible/playbooks/saas/roles/valkey/templates/nomad.hcl
index 55dc611d..026e2b06 100644
--- a/ansible/playbooks/saas/roles/valkey/templates/nomad.hcl
+++ b/ansible/playbooks/saas/roles/valkey/templates/nomad.hcl
@@ -3,7 +3,7 @@ job "{{ domain }}" {
   datacenters = ["{{ fact_instance.datacenter }}"]
   type = "service"
 
-{% if software.constraints.location %}
+{% if software.constraints is defined and software.constraints.location is defined %}
   constraint {
     attribute    = "${meta.location}"
     set_contains = "{{ software.constraints.location }}"
@@ -21,6 +21,9 @@ job "{{ domain }}" {
     network {
       port "valkey" {
         to = 6379
+{% if software.static_port is defined %}
+        static = {{ software.static_port }}
+{% endif %}
       }
     }
 
@@ -35,7 +38,7 @@ job "{{ domain }}" {
       driver = "docker"
 
       config {
-        image = "{{ software }}/{{ software }}:{{ softwares.valkey.version }}"
+        image = "valkey/valkey:{{ softwares.valkey.version }}"
         volumes = [
           "{{ software_path }}/data:/data:rw",
           "{{ software_path }}/etc/valkey:/etc/valkey:ro",
@@ -44,8 +47,16 @@ job "{{ domain }}" {
         ports = ["valkey"]
         command = "/usr/local/bin/valkey-server"
         args = [
-          "/etc/valkey/valkey.conf",
-]
+          "/local/valkey.conf",
+        ]
+      }
+
+      template {
+        change_mode = "restart"
+        destination = "local/valkey.conf"
+        data = <<EOH
+{{ lookup('ansible.builtin.template', 'templates/valkey.conf') }}
+  EOH
       }
 
       resources {
diff --git a/ansible/playbooks/saas/roles/valkey/files/valkey.conf b/ansible/playbooks/saas/roles/valkey/templates/valkey.conf
similarity index 99%
rename from ansible/playbooks/saas/roles/valkey/files/valkey.conf
rename to ansible/playbooks/saas/roles/valkey/templates/valkey.conf
index 62b42dea..d754a3aa 100644
--- a/ansible/playbooks/saas/roles/valkey/files/valkey.conf
+++ b/ansible/playbooks/saas/roles/valkey/templates/valkey.conf
@@ -85,7 +85,8 @@
 # You will also need to set a password unless you explicitly disable protected
 # mode.
 # ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
-bind 127.0.0.1 -::1
+# bind 127.0.0.1 -::1
+bind * -::*  
 
 # By default, outgoing connections (from replica to primary, from Sentinel to
 # instances, cluster bus, etc.) are not bound to a specific local address. In
@@ -109,7 +110,7 @@ bind 127.0.0.1 -::1
 # By default protected mode is enabled. You should disable it only if
 # you are sure you want clients from other hosts to connect to the server
 # even if no authentication is configured.
-protected-mode yes
+protected-mode no
 
 # The server uses default hardened security configuration directives to reduce the
 # attack surface on innocent users. Therefore, several sensitive configuration
diff --git a/ansible/playbooks/saas/roles/vector/README.md b/ansible/playbooks/saas/roles/vector/README.md
new file mode 100644
index 00000000..86fc9973
--- /dev/null
+++ b/ansible/playbooks/saas/roles/vector/README.md
@@ -0,0 +1 @@
+# Role: `vector`
diff --git a/ansible/playbooks/saas/roles/vector/defaults/main.yml b/ansible/playbooks/saas/roles/vector/defaults/main.yml
new file mode 100644
index 00000000..ed97d539
--- /dev/null
+++ b/ansible/playbooks/saas/roles/vector/defaults/main.yml
@@ -0,0 +1 @@
+---
diff --git a/ansible/playbooks/saas/roles/vector/tasks/backup.yml b/ansible/playbooks/saas/roles/vector/tasks/backup.yml
new file mode 100644
index 00000000..ed97d539
--- /dev/null
+++ b/ansible/playbooks/saas/roles/vector/tasks/backup.yml
@@ -0,0 +1 @@
+---
diff --git a/ansible/playbooks/saas/roles/vector/tasks/build.yml b/ansible/playbooks/saas/roles/vector/tasks/build.yml
new file mode 100644
index 00000000..7421b506
--- /dev/null
+++ b/ansible/playbooks/saas/roles/vector/tasks/build.yml
@@ -0,0 +1,14 @@
+---
+- name: Include upstream variables
+  ansible.builtin.include_vars: upstream.yml
+
+- name: Set custom variables
+  ansible.builtin.set_fact:
+    image_version: "{{ latest_version }}"
+    image_name: "{{ image.name }}"
+    image_labels: "{{ image.labels }}"
+    image_build: "{{ image.build }}"
+
+- name: End playbook if no new version
+  ansible.builtin.meta: end_host
+  when: softwares[image.name] is defined and softwares[image.name] == image_version
diff --git a/ansible/playbooks/saas/roles/vector/tasks/destroy.yml b/ansible/playbooks/saas/roles/vector/tasks/destroy.yml
new file mode 100644
index 00000000..e00b7de1
--- /dev/null
+++ b/ansible/playbooks/saas/roles/vector/tasks/destroy.yml
@@ -0,0 +1,11 @@
+---
+- name: Stop nomad job
+  ansible.builtin.include_role:
+    name: nomad
+    tasks_from: job_stop.yml
+
+- name: Remove software directory
+  ansible.builtin.file:
+    path: "{{ software_path }}"
+    state: absent
+  delegate_to: "{{ software.instance }}"
\ No newline at end of file
diff --git a/ansible/playbooks/saas/roles/vector/tasks/main.yml b/ansible/playbooks/saas/roles/vector/tasks/main.yml
new file mode 100644
index 00000000..148c7a41
--- /dev/null
+++ b/ansible/playbooks/saas/roles/vector/tasks/main.yml
@@ -0,0 +1,14 @@
+---
+- name: Copy nomad job
+  ansible.builtin.template:
+    src: nomad.hcl
+    dest: "/var/tmp/{{ domain }}.nomad"
+    owner: root
+    group: root
+    mode: '0600'
+  become: true
+
+- name: Run nomad job
+  ansible.builtin.include_role:
+    name: nomad
+    tasks_from: job_run.yml
diff --git a/ansible/playbooks/saas/roles/vector/tasks/restore.yml b/ansible/playbooks/saas/roles/vector/tasks/restore.yml
new file mode 100644
index 00000000..ed97d539
--- /dev/null
+++ b/ansible/playbooks/saas/roles/vector/tasks/restore.yml
@@ -0,0 +1 @@
+---
diff --git a/ansible/playbooks/saas/roles/vector/templates/nomad.hcl b/ansible/playbooks/saas/roles/vector/templates/nomad.hcl
new file mode 100644
index 00000000..ed39381a
--- /dev/null
+++ b/ansible/playbooks/saas/roles/vector/templates/nomad.hcl
@@ -0,0 +1,75 @@
+job "{{ domain }}" {
+  region = "{{ fact_instance.region }}"
+  datacenters = ["{{ fact_instance.datacenter }}"]
+  type = "service"
+
+{% if software.constraints is defined and software.constraints.location is defined %}
+  constraint {
+    attribute    = "${meta.location}"
+    set_contains = "{{ software.constraints.location }}"
+  }
+{% endif %}
+
+  constraint {
+    attribute    = "${meta.instance}"
+    set_contains = "{{ software.instance }}"
+  }
+
+  group "{{ domain }}" {
+    count = 1
+
+    network {
+      port "vector" {
+        to = 8686
+{% if software.static_port is defined %}
+        static = {{ software.static_port }}
+{% endif %}
+      }
+      port "vector_http" {
+        to = 8687
+        static = 8687
+      }
+    }
+
+    // service name
+    service {
+      name = "{{ service_name }}"
+      port = "vector"
+      provider = "nomad"
+      tags = [
+        {{ lookup('template', '../../traefik/templates/traefik_tag.j2') | indent(8) }}
+      ]
+    }
+
+    task "{{ domain }}-vector" {
+      driver = "docker"
+
+      template {
+        change_mode = "restart"
+        destination = "local/vector.yaml"
+        perms = "644"
+        data = <<EOH
+{{ lookup('ansible.builtin.template', 'templates/vector.yaml.j2') }}
+  EOH
+      }
+
+      config {
+        image = "timberio/vector:{{ softwares.vector.version }}-alpine"
+        volumes = [
+          "{{ software_path }}/var/lib/vector:/var/lib/vector:rw"
+        ]
+        ports = ["vector", "vector_http"]
+        args  = [
+          "--config",
+          "/local/vector.yaml"
+        ]
+      }
+
+      resources {
+        cpu    = {{ size[software.size].cpu }}
+        memory = {{ size[software.size].memory }}
+        memory_max = {{ size[software.size].memory * 2 }}
+      }
+    }
+  }
+}
diff --git a/ansible/playbooks/saas/roles/vector/templates/vector.yaml.j2 b/ansible/playbooks/saas/roles/vector/templates/vector.yaml.j2
new file mode 100644
index 00000000..27332fff
--- /dev/null
+++ b/ansible/playbooks/saas/roles/vector/templates/vector.yaml.j2
@@ -0,0 +1,25 @@
+data_dir: /var/lib/vector
+
+api:
+  enabled: true
+  address: 0.0.0.0:8686
+
+sources:
+  http:
+    type: http_server
+    address: 0.0.0.0:8687
+
+# sources:
+#   demo_logs:
+#     type: demo_logs
+#     interval: 1
+#     format: json
+
+sinks:
+  console:
+    inputs:
+      - http
+    target: stdout
+    type: console
+    encoding:
+      codec: json
\ No newline at end of file
diff --git a/ansible/playbooks/saas/roles/vector/vars/main.yml b/ansible/playbooks/saas/roles/vector/vars/main.yml
new file mode 100644
index 00000000..2b2ca167
--- /dev/null
+++ b/ansible/playbooks/saas/roles/vector/vars/main.yml
@@ -0,0 +1,13 @@
+---
+image:
+  build: false
+  upstream:
+    source: github
+    user: vectordotdev
+    repo: vector
+    type: releases
+    format: tar.gz
+    file: vector-VERSION-ARCH-unknown-OS-gnu.FORMAT
+    os: linux
+  labels: {}
+  name: vector
diff --git a/ansible/playbooks/saas/roles/vector/vars/upstream.yml b/ansible/playbooks/saas/roles/vector/vars/upstream.yml
new file mode 100644
index 00000000..03927468
--- /dev/null
+++ b/ansible/playbooks/saas/roles/vector/vars/upstream.yml
@@ -0,0 +1,3 @@
+---
+upstream_versions: "{{ (lookup('url', 'https://api.github.com/repos/{{ image.upstream.user }}/{{ image.upstream.repo }}/tags', headers={'Accept': 'application/vnd.github+json', 'Authorization': 'Bearer ' + lookup('ansible.builtin.env', 'GITHUB_API_TOKEN') })) }}"
+latest_version: "{{ upstream_versions | community.general.json_query('[].name') | select('match', '^(v|r)?\\d{1,4}(\\.\\d{1,4})(\\.\\d{1,4})$') | community.general.version_sort | last | replace('v', '') }}"
diff --git a/ansible/playbooks/saas/roles/vllm/README.md b/ansible/playbooks/saas/roles/vllm/README.md
new file mode 100644
index 00000000..371c3efa
--- /dev/null
+++ b/ansible/playbooks/saas/roles/vllm/README.md
@@ -0,0 +1 @@
+# Role: `vllm`
diff --git a/ansible/playbooks/saas/roles/vllm/defaults/main.yml b/ansible/playbooks/saas/roles/vllm/defaults/main.yml
new file mode 100644
index 00000000..ed97d539
--- /dev/null
+++ b/ansible/playbooks/saas/roles/vllm/defaults/main.yml
@@ -0,0 +1 @@
+---
diff --git a/ansible/playbooks/saas/roles/vllm/tasks/backup.yml b/ansible/playbooks/saas/roles/vllm/tasks/backup.yml
new file mode 100644
index 00000000..ed97d539
--- /dev/null
+++ b/ansible/playbooks/saas/roles/vllm/tasks/backup.yml
@@ -0,0 +1 @@
+---
diff --git a/ansible/playbooks/saas/roles/vllm/tasks/build.yml b/ansible/playbooks/saas/roles/vllm/tasks/build.yml
new file mode 100644
index 00000000..7421b506
--- /dev/null
+++ b/ansible/playbooks/saas/roles/vllm/tasks/build.yml
@@ -0,0 +1,14 @@
+---
+- name: Include upstream variables
+  ansible.builtin.include_vars: upstream.yml
+
+- name: Set custom variables
+  ansible.builtin.set_fact:
+    image_version: "{{ latest_version }}"
+    image_name: "{{ image.name }}"
+    image_labels: "{{ image.labels }}"
+    image_build: "{{ image.build }}"
+
+- name: End playbook if no new version
+  ansible.builtin.meta: end_host
+  when: softwares[image.name] is defined and softwares[image.name] == image_version
diff --git a/ansible/playbooks/saas/roles/vllm/tasks/destroy.yml b/ansible/playbooks/saas/roles/vllm/tasks/destroy.yml
new file mode 100644
index 00000000..ce77a12b
--- /dev/null
+++ b/ansible/playbooks/saas/roles/vllm/tasks/destroy.yml
@@ -0,0 +1,10 @@
+---
+- name: Stop nomad job
+  ansible.builtin.include_role:
+    name: nomad
+    tasks_from: job_stop.yml
+
+- name: Remove software directory
+  ansible.builtin.file:
+    path: "{{ software_path }}"
+    state: absent
diff --git a/ansible/playbooks/saas/roles/vllm/tasks/main.yml b/ansible/playbooks/saas/roles/vllm/tasks/main.yml
new file mode 100644
index 00000000..87410702
--- /dev/null
+++ b/ansible/playbooks/saas/roles/vllm/tasks/main.yml
@@ -0,0 +1,60 @@
+---
+- name: Copy llm nomad job to destination
+  ansible.builtin.template:
+    src: nomad-llm.hcl
+    dest: "/var/tmp/{{ domain }}.nomad"
+    owner: root
+    group: root
+    mode: '0600'
+  become: true
+
+- name: Run nomad llm job
+  ansible.builtin.include_role:
+    name: nomad
+    tasks_from: job_run.yml
+
+- name: Copy nomad job to destination
+  ansible.builtin.template:
+    src: nomad.hcl
+    dest: "/var/tmp/{{ domain }}.nomad"
+    owner: root
+    group: root
+    mode: '0600'
+  become: true
+
+- name: Run nomad job
+  ansible.builtin.include_role:
+    name: nomad
+    tasks_from: job_run.yml
+
+
+- name: Check for endpoint to become available
+  ansible.builtin.uri:
+    url: "https://{{ software.domain }}/v1/models"
+  delegate_to: localhost
+  register: _result
+  until: _result.status == 200
+
+- name: Warm up model with a simple query
+  ansible.builtin.uri:
+    url: https://{{ software.domain }}/v1/chat/completions
+    method: POST
+    body_format: json
+    body:
+      model: "{{ software.model }}"
+      messages:
+        - role: user
+          content: hello
+      max_tokens: 50
+    headers:
+      Content-Type: "application/json"
+    return_content: true
+  register: _result
+  until: _result.status == 200
+  retries: 10
+  delay: 10
+  delegate_to: localhost
+
+- name: Afficher la réponse du modèle
+  ansible.builtin.debug:
+    msg: "{{ _result.json }}"
diff --git a/ansible/playbooks/saas/roles/vllm/tasks/restore.yml b/ansible/playbooks/saas/roles/vllm/tasks/restore.yml
new file mode 100644
index 00000000..ed97d539
--- /dev/null
+++ b/ansible/playbooks/saas/roles/vllm/tasks/restore.yml
@@ -0,0 +1 @@
+---
diff --git a/ansible/playbooks/saas/roles/vllm/templates/entrypoint-llm.sh b/ansible/playbooks/saas/roles/vllm/templates/entrypoint-llm.sh
new file mode 100644
index 00000000..a4d522d8
--- /dev/null
+++ b/ansible/playbooks/saas/roles/vllm/templates/entrypoint-llm.sh
@@ -0,0 +1,10 @@
+#!/bin/sh
+
+apk update
+apk add --no-cache python3 py3-pip
+pip install  --break-system-packages -U "huggingface_hub[cli]" hf-transfer
+
+mkdir -p /models/${MODEL}
+
+echo "Download model ..."
+hf download meta-llama/${MODEL} --local-dir /models/${MODEL}
\ No newline at end of file
diff --git a/ansible/playbooks/saas/roles/vllm/templates/nomad-llm.hcl b/ansible/playbooks/saas/roles/vllm/templates/nomad-llm.hcl
new file mode 100644
index 00000000..a69885ae
--- /dev/null
+++ b/ansible/playbooks/saas/roles/vllm/templates/nomad-llm.hcl
@@ -0,0 +1,67 @@
+job "{{ domain }}-llm" {
+  region = "{{ fact_instance.region }}"
+  datacenters = ["{{ fact_instance.datacenter }}"]
+  type = "batch"
+
+{% if software.constraints is defined and software.constraints.location is defined %}
+  constraint {
+    attribute    = "${meta.location}"
+    set_contains = "{{ software.constraints.location }}"
+  }
+{% endif %}
+
+{% if software.constraints is defined and software.constraints.instance is defined %}
+  constraint {
+    attribute    = "${meta.instance}"
+    set_contains = "{{ software.constraints.instance }}"
+  }
+{% endif %}
+
+  group "{{ domain }}-llm" {
+    count = 1
+
+    volume "models" {
+      type            = "csi"
+      source          = "juicefs-llm"
+      attachment_mode = "file-system"
+      access_mode     = "multi-node-multi-writer"
+    }
+
+    task "{{ domain }}-vllm" {
+      driver = "docker"
+
+      template {
+        change_mode = "restart"
+        destination = "local/entrypoint.sh"
+        perms = "755"
+        data = <<EOH
+{{ lookup('ansible.builtin.template', 'templates/entrypoint-llm.sh') }}
+  EOH
+      }
+
+      env {
+        HF_HOME = "/models"
+        HF_TOKEN = "{{ software.hf_token }}"
+        HF_HUB_DOWNLOAD_TIMEOUT = "60"
+        HF_HUB_DISABLE_XET = "1"
+        HF_HUB_ENABLE_HF_TRANSFER = "1"
+        MODEL = "{{ software.model }}"
+      }
+
+      config {
+        image = "alpine:latest"
+        command = "/local/entrypoint.sh"
+      }
+
+      volume_mount {
+        volume      = "models"
+        destination = "/models"
+      }
+
+      resources {
+        cpu    = {{ size[software.size].cpu }}
+        memory = {{ size[software.size].memory }}
+      }
+    }
+  }
+}
diff --git a/ansible/playbooks/saas/roles/vllm/templates/nomad.hcl b/ansible/playbooks/saas/roles/vllm/templates/nomad.hcl
new file mode 100644
index 00000000..da4017b0
--- /dev/null
+++ b/ansible/playbooks/saas/roles/vllm/templates/nomad.hcl
@@ -0,0 +1,115 @@
+job "{{ domain }}" {
+  region = "{{ fact_instance.region }}"
+  datacenters = ["{{ fact_instance.datacenter }}"]
+  type = "service"
+
+{% if software.constraints is defined and software.constraints.location is defined %}
+  constraint {
+    attribute    = "${meta.location}"
+    set_contains = "{{ software.constraints.location }}"
+  }
+{% endif %}
+
+{% if software.constraints is defined and software.constraints.instance is defined %}
+  constraint {
+    attribute    = "${meta.instance}"
+    set_contains = "{{ software.constraints.instance }}"
+  }
+{% endif %}
+
+  group "{{ domain }}" {
+    count = {{ software.scale | default(1) }}
+
+    constraint {
+      operator  = "distinct_hosts"
+      value     = "true"
+    }
+
+    network {
+      port "vllm" {
+        to = 8000
+      }
+    }
+
+    volume "models" {
+      type            = "csi"
+      source          = "juicefs-llm"
+      attachment_mode = "file-system"
+      access_mode     = "multi-node-multi-writer"
+    }
+
+    service {
+      name = "{{ service_name }}"
+      port = "vllm"
+      provider = "nomad"
+      tags = [
+        {{ lookup('template', '../../traefik/templates/traefik_tag.j2') | indent(8) }}
+      ]
+      check {
+        type     = "http"
+        path     = "/v1/models"
+        interval = "10s"
+        timeout  = "2s"
+      }
+    }
+
+    service {
+      name = "vllm"
+      port = "vllm"
+      provider = "nomad"
+      tags = [
+        "fqdn:{{ domain }}",
+        "host:{{ inventory_hostname }}",
+      ]
+    }
+
+    task "{{ domain }}-vllm" {
+      driver = "docker"
+
+      env {
+        HF_HOME = "/models"
+        HF_TOKEN = "{{ software.hf_token }}"
+        HUGGING_FACE_HUB_TOKEN = "{{ software.hf_token }}"
+        VLLM_USE_V1 = "1"
+        #VLLM_ALLREDUCE_USE_SYMM_MEM = "0"
+        VLLM_DO_NOT_TRACK = "0"
+        VLLM_USE_FLASHINFER_SAMPLER = "0"
+        TORCH_CUDA_ARCH_LIST = "Ada"
+        CUDA_VISIBLE_DEVICES = "0"
+      }
+
+      config {
+        image = "vllm/vllm-openai:v{{ software.image_version | default(softwares.vllm.version) }}"
+        ports = ["vllm"]
+        ipc_mode = "host"
+
+        volumes = [
+          "{{ software_path }}/.config:/root/.config:rw",
+          "{{ software_path }}/.cache:/root/.cache:rw"
+        ]
+        args = [{% for arg in software.args %}"{{ arg }}"{% if not loop.last %},{% endif %}{% endfor %}]
+        work_dir = "/models"
+      }
+
+      volume_mount {
+        volume      = "models"
+        destination = "/models"
+      }
+
+      resources {
+        cpu    = {{ size[software.size2 | default(software.size)].cpu }}
+        memory = {{ size[software.size2 | default(software.size)].memory }}
+        
+        device "nvidia/gpu" {
+          count = 1
+          affinity {
+            attribute = "${device.model}"
+            operator  = "regexp"
+            value     = "Tesla"
+          }
+        }
+      }
+
+    }
+  }
+}
diff --git a/ansible/playbooks/saas/roles/vllm/vars/main.yml b/ansible/playbooks/saas/roles/vllm/vars/main.yml
new file mode 100644
index 00000000..16313900
--- /dev/null
+++ b/ansible/playbooks/saas/roles/vllm/vars/main.yml
@@ -0,0 +1,13 @@
+---
+image:
+  build: false
+  upstream:
+    source: github
+    user: vllm-project
+    repo: vllm
+    type: release
+    format: tar.gz
+    file: vllm-VERSION.FORMAT
+  labels: {}
+  name: vllm
+  origin: ubuntu:latest
diff --git a/ansible/playbooks/saas/roles/vllm/vars/upstream.yml b/ansible/playbooks/saas/roles/vllm/vars/upstream.yml
new file mode 100644
index 00000000..036aa64c
--- /dev/null
+++ b/ansible/playbooks/saas/roles/vllm/vars/upstream.yml
@@ -0,0 +1,4 @@
+---
+latest_version: "{{ (lookup('url', 'https://api.github.com/repos/{{ image.upstream.user }}/{{ image.upstream.repo }}/releases/latest', headers={'Accept': 'application/vnd.github+json', 'Authorization': 'Bearer ' + lookup('ansible.builtin.env', 'GITHUB_API_TOKEN') }) | from_json).get('tag_name') | replace('v', '') }}"
+upstream_file_name: "{{ image.upstream.file | replace('REPO', image.upstream.repo) | replace('VERSION', latest_version) | replace('OS', image.upstream.os) | replace('ARCH', upstream_default_arch) | replace('FORMAT', image.upstream.format) }}"
+upstream_file_url: "https://github.com/{{ image.upstream.user }}/{{ image.upstream.repo }}/releases/download/v{{ latest_version }}/{{ upstream_file_name }}"
diff --git a/ansible/playbooks/saas/roles/wordpress/templates/nomad.hcl b/ansible/playbooks/saas/roles/wordpress/templates/nomad.hcl
index ab54a4e8..2e615550 100644
--- a/ansible/playbooks/saas/roles/wordpress/templates/nomad.hcl
+++ b/ansible/playbooks/saas/roles/wordpress/templates/nomad.hcl
@@ -3,7 +3,7 @@ job "{{ domain }}" {
   datacenters = ["{{ fact_instance.datacenter }}"]
   type        = "service"
 
-{% if software.constraints.location %}
+{% if software.constraints is defined and software.constraints.location is defined %}
   constraint {
     attribute    = "${meta.location}"
     set_contains = "{{ software.constraints.location }}"
diff --git a/ansible/playbooks/saas/roles/zigbee2mqtt/templates/nomad.hcl b/ansible/playbooks/saas/roles/zigbee2mqtt/templates/nomad.hcl
index 255ef6d9..4cf709e6 100644
--- a/ansible/playbooks/saas/roles/zigbee2mqtt/templates/nomad.hcl
+++ b/ansible/playbooks/saas/roles/zigbee2mqtt/templates/nomad.hcl
@@ -3,7 +3,7 @@ job "{{ domain }}" {
   datacenters = ["{{ fact_instance.datacenter }}"]
   type = "service"
 
-{% if software.constraints.location %}
+{% if software.constraints is defined and software.constraints.location is defined %}
   constraint {
     attribute    = "${meta.location}"
     set_contains = "{{ software.constraints.location }}"
diff --git a/ui/controllers/api.js b/ui/controllers/api.js
index d7c3d3ae..d75d2620 100644
--- a/ui/controllers/api.js
+++ b/ui/controllers/api.js
@@ -51,11 +51,12 @@ exports.install = function() {
 	ROUTE('+API    /api/       +softwares_execute/{id}               --> Softwares/execute');
 	
 	// Variables
-	ROUTE('+API    /api/       +variables		          --> Variables/list');
-	ROUTE('+API    /api/       +variables_read/{id}      --> Variables/read');
-	ROUTE('+API    /api/       +variables_create          --> Variables/create');
-	ROUTE('+API    /api/       +variables_update/{id}    --> Variables/update');
-	ROUTE('+POST   /api/secret                         --> Variables/secret');
+	ROUTE('+API    /api/       +variables		                     --> Variables/list');
+	ROUTE('+API    /api/       +variables_read/{id}                  --> Variables/read');
+	ROUTE('+API    /api/       +variables_create                     --> Variables/create');
+	ROUTE('+API    /api/       +variables_update/{id}                --> Variables/update');
+	ROUTE('+API    /api/       +variables_remove/{id}                --> Variables/remove');
+	ROUTE('+POST   /api/secret                                       --> Variables/secret');
 
 	// 3dForceGraph
 	ROUTE('+API    /api/       -graphs					  --> Graphs/list');
diff --git a/ui/public/forms/catalogs.html b/ui/public/forms/catalogs.html
index c00feb78..bad80e7c 100644
--- a/ui/public/forms/catalogs.html
+++ b/ui/public/forms/catalogs.html
@@ -17,6 +17,7 @@
 <ui-component name="box" path="common.form" config="if:~PATH~;width:620;icon:ti ti-folders;title:@(Catalog);reload:?/reload;" class="hidden ~PATH~" plugin="~PATH~">
 	<nav>
 		<button class="exec" data-exec="?/settings"><i class="ti ti-config"></i>@(Settings)</button>
+		<button class="exec" data-exec="?/add"><i class="ti ti-plus-circle green"></i>@(Add)</button>
 	</nav>
 	<ui-component name="textbox" path="?.search" config="type:search"></ui-component>
 	<ui-component name="searchdata" path="?.search" config="datasource:?.items;key:name;output:?.searchoutput">
@@ -70,6 +71,11 @@
 			});
 		};
 
+		exports.add = function() {
+			SET('formcatalog @default', {});
+			SET('common.form2', 'formcatalog');
+		};
+
 		exports.settings = function(el) {
 			const id = GET('?.settingsId');
 			const type = 'settings';
diff --git a/ui/public/forms/variables.html b/ui/public/forms/variables.html
index 4ce55557..8b344864 100644
--- a/ui/public/forms/variables.html
+++ b/ui/public/forms/variables.html
@@ -25,9 +25,9 @@
 				{{ foreach m in value }}
 					<figure data-id="{{ m.id }}" data-key="{{ m.key }}" data-type="{{ m.type }}" class="exec items" data-exec="?/update">
 						<section>
-							<!-- <div class="buttons">
-								<span class="badge" data-prevent="true">{{ m.type }}</span>
-							</div> -->
+							<div class="buttons">
+								<span class="exec" data-exec="?/remove" data-prevent="true"><i class="ti ti-remove red"></i></span>
+							</div>
 							<div class="name">
 								<div class="picto">
 									<i class="ti ti-{{ m.type | type }}" style="color:{{ m.type | color }}"></i>
@@ -128,5 +128,12 @@
 				SET('common.form2', 'formvariable');
 			});
 		};
+
+		exports.remove = function(el) {
+			var id = ATTRD(el);
+			EXEC('-approve/show', '@(Remove selected variable?)', '"ti ti-remove" @(Remove)', function() {
+				exports.tapi('variables_remove/{0} ERROR'.format(id), exports.refresh);
+			});
+		};
 	});
 </script>
\ No newline at end of file
diff --git a/ui/public/img/generic.png b/ui/public/img/generic.png
new file mode 100644
index 00000000..63d3282f
Binary files /dev/null and b/ui/public/img/generic.png differ
diff --git a/ui/public/img/vllm.png b/ui/public/img/vllm.png
new file mode 100644
index 00000000..6948455e
Binary files /dev/null and b/ui/public/img/vllm.png differ
diff --git a/ui/schemas/softwares.js b/ui/schemas/softwares.js
index 7cbf9e16..47763368 100644
--- a/ui/schemas/softwares.js
+++ b/ui/schemas/softwares.js
@@ -44,24 +44,20 @@ NEWSCHEMA('Softwares', function (schema) {
 		name: 'Search item',
 		params: '*prefix:String,*action:String,*api:String, alias:String',
 		action: async function ($) {
-			try {
-				const { prefix, action, api, alias } = $.params;
-				const list = await $.action('Softwares/list').promise($);
-				const result = list.map(item => ({
-					id: action,
-					api,
-					form: 'formsoftware',
-					search: item.domain,
-					name: item.domain,
-					itemid: item.id,
-					html: `${prefix}: <span class="gray">${alias || action} ${item.domain}</span>`,
-					icon: 'ti-cube',
-					color: 'blue'
-				}));
-				$.callback(result);
-			} catch (err) {
-				$.error(err);
-			}
+			const { prefix, action, api, alias } = $.params;
+			const list = await $.action('Softwares/list').promise($);
+			const result = list.map(item => ({
+				id: action,
+				api,
+				form: 'formsoftware',
+				search: item.domain,
+				name: item.domain,
+				itemid: item.id,
+				html: `${prefix}: <span class="gray">${alias || action} ${item.domain}</span>`,
+				icon: 'ti-cube',
+				color: 'blue'
+			}));
+			$.callback(result);
 		}
 	});