Moqui Metrics

A module for moqui-framework to make metrics available to monitoring tools, like Prometheus.

Sample Grafana Dashboard

See sample Grafana Dashboard in file grafana/MoquiDashboard.json, which includes visualizations like following:

Restart-OOM and Throughput

Resource Usage

Memory Visualizations

Complete for several namespaces

And filtered by specific namespace

Artifact Versions

See which specific commits were running in a specific pod at a given time.

Supported Monitoring Tools

Currently, only Prometheus is supported. The URL to be used as target is the /metrics/prometheus path of your publically reachable moqui instance. If in your configuration the public URL might reach more than one instance, you should add a unique name for each instance to be reached consistently, by e.g. using additional host names that are mapped to only one instance.

Configuration

This module currently has a setting to configure allowing access based on IP addresses using the metrics_prometheus_servers property. This property can be set using java properties, environment vars or direct configuration in a MoquiConf.xml file. The value is a comma-separated list of hosts (hostname or IP address), or the special value private which will match all IPv4 private address classes.

Exporting metrics for Artifacts

Metrics for Artifacts are exported using the prometheus histogram type. You can create different histogram groups by adding values to the moqui.metrics.MetricsHitGroup entity, and associating one or more artifacts to this counter using the moqui.metrics.MetricsHitGroupArtifact entry. By default, for durations of 5, 10, 25, 50, 100, 250, 500, 1000, 3000, 5000 and 10000 ms are created, but you may specify custom buckets using the moqui.metrics.MetricsHitGroupBucket entity.

The values for artifactType and artifactName should be the same than used in the moqui.server.ArtifactHit entity.

For example, in order to export all accesses to the Tools app, the artifactType would be set to AT_XML_SCREEN and the artifactName to component://tools/screen/Tools.xml See MetricsHitGroupDemoData for details.

Prometheus

Standalone Usage

1. Set the address(es) of the prometheus server as specified in the configuration
2. Configure prometheus to scrape the corresponding URL. For a prometheus server with a static configuration, something like this:

    scrape_configs:
      - job_name: 'External Moqui Instances'
        metrics_path: '/metrics/prometheus'
        scheme: https
        static_configs:
          - targets: [instance1.my-domain.com]
            labels:
              instance: 'instance1'
          - targets: [instance2.my-domain.com]
            labels:
              instance: 'instance1'
   

Kubernetes Usage

Scraping auto-discovery

1. Make sure your prometheus server has endpoint services auto-discovery enabled (default for most installations)
2. Add the auto-discover labels to your moqui pods:
   • prometheus.io/scrape=true
   • prometheus.io/path=/metrics/prometheus
   • prometheus.io/port=80 (or other port like 8080, according to your pod mapping)

Pod Monitor

Configure a PodMonitor that matches the running instance(s), e.g.:

apiVersion: monitoring.coreos.com/v1
kind: PodMonitor
spec:
  namespaceSelector:
    matchNames:
      - moit-dte
  podMetricsEndpoints:
    - path: /metrics/prometheus
      port: http
      scheme: http
  selector:
    matchLabels:
      app: moqui