Use case

Planned

Incident postmortem

Run a postmortem as governed stages, and keep the lessons where the next decision can find them.

What it is

The stage, and what it produces

A postmortem is reviewed, approved work whose whole point is to be remembered. The governed-run pattern fits it well: produce the analysis, have a human approve it, and keep the lessons in Memory so they inform future decisions.

In a run

How it works, step by step

01

Gather the timeline and contributing factors

02

Propose findings and actions for review

03

A human approves the conclusions

04

Lessons kept in Memory for next time

What caused the double-charge?
  • Timeline · Retry fired before the ack
  • Cause · Missing idempotency key
  • Action · Enforce key on /charges

Illustrative — produced, reviewed and approved before the next step.

Works with

The parts of the platform it leans on

Where it stands. Exploratory. The Memory that would keep the lessons ships today; a postmortem-specific workflow is not something we've shipped.

Turn this stage into a governed output

Start with what ships today, and grow into the full run as it lands.