The blameless post mortem: how embracing failure makes us better
summary
Argues that blameless post‑mortems turn incidents into prevention by prioritizing fixes over finger‑pointing, using NTSB and M&M techniques like fishbone diagrams and a neutral chairperson. Gives a concrete incident where deleted the github pages branch causes two hours without docs, detection lag ~30 minutes and a 50 minute publish delay; fixes include enable branch protection, better logging, and considering alternatives to GitHub Pages. Emphasizes psychological safety as necessary for honest reports and actionable prevention.
AI-written summary · about
chapters (32)
- 0:00Intro
- 1:04Website evolution
- 1:29What if
- 4:23VB Will
- 4:41Culture
- 5:26NTSB
- 6:41Sharing faults
- 7:33NTSB report
- 8:03How did we get here
- 9:32MM conferences
- 12:14Who should be present
- 13:52Fishbone diagram
- 14:24Words change
- 15:05What can we learn
- 15:52Three components
- 16:05Learning from failure
- 16:42Example incident
- 19:51Timeline
- 20:41Luck
- 22:35Action Items
- 23:49Fail Fast
- 24:29Deterrence
- 25:19Excuses
- 25:45Example
- 26:41Create psychological safety
- 26:55Rework from Google
- 27:29Site Reliability Engineering
- 28:09Outages will always happen
- 28:23A test for blamelessness
- 29:52A quote from an economist
- 30:05If youre never failing
- 30:16SLA
related talks
28mBlameless Postmortem Culture
23mWho to blame for all your problems
51mEmbracing Failure and Learning from Our Mistakes with Effective Post Mortems
16mInterview
29mLearning from incidents: how things went right
13mSS & BR: Post Mortem — what an autopsy can give us
22mBlame, shame and panic - how not to respond when things go wrong
31mPost-mortems: Building better software together.