Skip to main content

Validate Slack Output

Overview​

After connecting Slack, verify that investigation results are delivered correctly to your configured channel.

Validation Steps​

1. Trigger an Investigation​

There is no manual "test investigation" button. To produce a message, run a real investigation one of two ways:

  • Wait for a matching alert. When an alert from a connected source matches an alert rule with auto-investigation enabled, OpsWorker investigates and posts to the routed channel.
  • Use AI Chat. Start an investigation from chat for a namespace whose routing you configured, then check that channel.

2. Check Your Slack Channel​

Within 2 minutes, a message should appear in the Slack channel routed for that Kubernetes namespace. The message includes:

  • Alert summary: the alert that triggered the investigation
  • Root cause analysis: what the AI identified as the underlying issue
  • Affected resources: Kubernetes resources involved (pods, services, deployments)
  • Recommended actions: specific remediation steps with kubectl commands
  • Feedback buttons: three buttons to rate investigation quality

3. Test Feedback Buttons​

Click one of the feedback buttons on the Slack message:

ButtonMeaning
AccurateThe root cause and recommendations were correct
Partially AccurateSome findings were useful, but incomplete or partially wrong
Needs ImprovementThe investigation missed the mark

Feedback helps OpsWorker improve future investigation quality.

Troubleshooting​

If no message appears in Slack:

  1. Check Slack integration status: in the portal, verify the Slack integration shows as connected under Integrations
  2. Check namespace routing: verify the investigation's Kubernetes namespace is mapped to a channel in the Slack integration screen
  3. Check channel permissions: ensure the OpsWorker bot has been invited to the target channel (some workspaces require this)
  4. Check investigation status: verify the investigation completed successfully in the portal under Investigations
  5. Re-authorize: if the integration shows errors, disconnect and reconnect Slack

Expected Message Format​

A typical Slack investigation message looks like:

🔍 Investigation Complete: HighPodCPUUsage

Root Cause: Pod api-gateway in namespace production is experiencing
CPU throttling due to resource limits set below actual usage patterns.

Affected Resources:
• Pod: api-gateway-7d8f9b6c4-x2k9p (production)
• Deployment: api-gateway (production)

Recommended Actions:
1. Increase CPU limit: kubectl patch deployment api-gateway -n production ...
2. Review HPA configuration for auto-scaling

[Accurate] [Partially Accurate] [Needs Improvement]

Next Steps​