Slack-Initiated Deep Stack Diagnosis #4
Description
As a DevOps engineer,
I want to trigger deep stack diagnosis directly from a Slack alert,
So that I can understand root causes that span beyond a single pod without switching tools.
🔍 Acceptance Criteria
➤ Trigger
Slack alert includes an interactive button: 🧠 Run Deep Diagnosis
Clicking the button:
Sends signed Slack action payload to backend
Payload includes pod metadata and alert ID
➤ Execution
Backend:
Validates action signature + alert origin
Looks up pod in cluster
Determines Helm release (via helm ls or annotations)
Gathers metadata, logs, events from dependent services
Sends structured payload to LLM backend
Timeout after 30s if deep analysis takes too long
➤ Response
Posts result in Slack thread:
📊 Root cause chain (e.g., Redis OOM → worker fail → 500s)
🔧 Fixes per component
⏱ Time taken
🧠 Powered by Opsctrl AI (for branding)
➤ Limits & Quotas
Per-org usage cap (e.g., 3 per day)
If quota exceeded, disable button with message:
“Deep diagnosis quota reached. Try again tomorrow.”
Optional: Only enabled for Slack-configured orgs or premium plans