The blameless postmortem

A postmortem is a write-up of an incident: what happened, when, why, what made it better, what made it worse, and what will change. The word blameless is important. The goal is to stop the same incident from recurring, not to find someone to scold.

A good postmortem has six sections. Summary. Timeline of events. Root cause. Contributing factors. What went well. Action items with owners and dates. Every action item has a concrete change someone has committed to make: a test, a runbook, an alert, a code change.