59 / 81 · Module 16 — Production Engineering and Observability · filtered: interview← prev⊞ allnext →
16.46Runbooks
For common incidents, create documented procedures.
Example:
Salesforce Connectivity Failure Runbook
could say:
1. Confirm error rate.
2. Check Salesforce status/network.
3. Check connector auth.
4. Check recent deployments.
5. Check retry exhaustion.
6. Determine if requests are queued.
7. Do NOT manually replay non-idempotent operations.
8. Query affected records before replay.
Runbooks reduce random production improvisation.