Incident triage

Stage 1 · 2 agents in parallel

• LookoutObservability & incidentsQueuedRunning✓ Done

  Scope the symptom

  Which services, since when, how many requests. Reads the error rate, latency and restarts around the window.

• HelmsmanReleasesQueuedRunning✓ Done

  List what changed

  Deployments, config changes, secret rotations and scaling events in the same window.

Stage 2

• LookoutObservability & incidentsQueuedRunning✓ Done

  Read the evidence

  Pulls the relevant log lines from the affected service and its upstreams and matches them to a change.

Stage 3

• KaptenOrchestrationQueuedRunning✓ Done

  Likely cause and options

  Proposes a cause with its evidence, and ranked mitigations: roll back, scale, or toggle config.

Stage 4

• HelmsmanReleasesQueuedRunning✓ Done

  Mitigate

  Applies the mitigation you approve, then watches health until the error rate returns to baseline.

Stage 5

• KaptenOrchestrationQueuedRunning✓ Done

  Draft the postmortem

  A timeline from the first error to recovery, the cause and the follow-ups to schedule.