Using postmortems to break out of infinite loops
summary
A concrete story shows a read replica accidentally becoming primary, a long SELECT count() triggers a payments outage, and the postmortem evolves from blaming people to adding a checklist to actually automate revoke access on the primary DB. The clear claim is that effective postmortems map a tight root cause to action items* and prefer system safeguards over exhortations to “be careful,” with the final fix implemented and tested so the outage cannot recur.
AI-written summary · about
chapters (6)
- 0:00Postmortems and pitfalls
- 3:53What a postmortem means
- 6:21First draft flaws
- 7:23Refining root cause
- 8:47Building safeguards
- 10:14Key takeaways
related talks
28mBlameless Postmortem Culture
30mLearning From Failure: Post Mortems
16mInterview
07mDia de los Muertos - Postmortems that save lives
27mPostmortem Action Items: Plan the Work and Work the Plan
28mAI-Driven Incident Response: From Alert to Postmortem
31mPost-mortems: Building better software together.
40mPost-Mortem