Postmortem Structure and Action Items
What you'll learn: How to organize a postmortem document and define concrete follow-up tasks that turn incident learnings into reliability improvements.
The Core Structure
A well-structured postmortem transforms chaos into clarity. Think of it like a well-organized crime scene investigation report—it should tell a clear story and lead to actionable prevention.
Essential sections:
- Timeline: Chronological sequence of what happened, when, and who did what. Timestamps matter.
- Root Cause: The underlying reason the incident occurred (not just symptoms).
- Contributing Factors: Additional conditions that enabled or worsened the incident.
- Impact Metrics: Quantifiable damage—error rate, affected users, revenue loss, MTTR.
- Detection/Response Analysis: How was it detected? How long until response? What worked or didn't?
From Analysis to Action
The magic happens in the Action Items section. These aren't vague wishes—they're concrete tasks:
- Specific: "Add timeout to database query X" not "improve monitoring"
- Owned: Every action has a name attached
- Tracked: Added to your project management system with deadlines
- Prioritized: Critical fixes first, nice-to-haves later
Example action items:
- "Add circuit breaker to payment service (Owner: Sarah, Due: 2 weeks)"
- "Document rollback procedure for DB migrations (Owner: James, Due: 1 week)"
- "Set up canary deployment for API changes (Owner: Platform team, Due: Q2)"
The detection/response analysis often reveals process gaps: "We didn't have an on-call runbook" becomes "Create runbook for X scenario."
Key Takeaway: A postmortem without tracked, owned action items is just storytelling—the structure turns incidents into systematic reliability improvements that prevent recurrence.