Postmortem (Blameless Incident Review)

Turn an incident into learning and prevention — focused on systems and process, never on blaming individuals.

When to use

• After an outage, regression, data issue, security event, or significant miss.

For sprint/project reflection (not incidents) use retrospective.

Principles

• Blameless. Assume everyone acted reasonably with the information they had. Fix systems, not people. Replace "X made a mistake" with "the system allowed X to happen."
• Facts first. Build the timeline from evidence (logs, alerts, messages) before drawing conclusions.
• Root cause, not first cause. Use the 5 Whys / contributing-factors analysis to get past the proximate trigger.
• Actions must prevent recurrence, be owned, and be dated.

Inputs to gather

• What happened, when detected, when resolved; severity and who/what was impacted.
• The timeline of events and the response.
• Logs/notes if available.

Process

1. Write a summary: what happened, impact (users/duration/scope), severity.
2. Build a timeline (UTC or stated TZ): detection → diagnosis → mitigation → resolution. Note time-to-detect and time-to-resolve.
3. Analyze root cause(s) and contributing factors (5 Whys).
4. Note what went well and what made it worse (detection gaps, slow rollback, unclear ownership).
5. Define corrective actions: preventive + detective, each with owner, due date, priority.
6. Capture lessons learned.

Output format

# Postmortem — {{incident}} · {{date}}
**Severity:** {{SEV}} · **Impact:** {{users/scope/duration}} · **Status:** Resolved

## Summary
{{2–4 sentences}}

## Timeline
| Time | Event |

## Root cause & contributing factors
{{5 Whys / analysis}}

## What went well / what went poorly
- ...

## Action items
| Action | Type (prevent/detect) | Owner | Due | Priority |

Keep tone neutral and factual throughout.