🚨

Incident Postmortem Questions

30 items

Incident reviews drift toward who rather than why within about ten minutes, and once they do nobody volunteers anything useful again. Structured questions hold the line. Draw five or six and work through them in order. Thirty prompts covering detection, response, contributing factors and the honest ones about what made the system easy to break. Note this is distinct from a sprint retro. An incident has a timeline, and the timeline is the material.

What's on the Wheel

All 30 entries that will be loaded onto the spinning wheel

1What was the impact, in customer terms?
2When did it actually start?
3When did we first know?
4How did we find out?
5Should we have found out sooner?
6What alerted, and what should have?
7What was the first thing we tried?
8What did we believe was happening that was wrong?
9What made diagnosis slow?
10What made diagnosis fast?
11Who did we need and could not reach?
12What documentation was missing or wrong?
13What did we have to guess at?
14What made the fix risky?
15How did we verify recovery?
16What contributed that was not the direct cause?
17What made this failure possible in the first place?
18Had this nearly happened before?
19What warnings did we not act on?
20What would have limited the blast radius?
21What would have made it self-heal?
22What manual step should be automated?
23What would a new joiner have struggled with?
24Was the on-call rota workable during this?
25How was the communication to customers?
26What did we learn that surprises us?
27What are we changing, specifically?
28Who owns each action and by when?
29What are we deliberately not changing?
30How will we know the fix worked?
🚨

Ready to spin?

This list is pre-loaded and ready to go. Click below to open the wheel and spin.

Spin the Wheel Now