Post Mortems

TL;DR - A Post Mortem is a reflection / review of an incident to better understand what happened, the scenario that allowed it to happen, and how to prevent it from recurring.

Overview

Post Mortems are a useful practice of reflection for understanding how an incident occurred and coming up with next steps to prevent its recurrence.

Properties of a good Post Mortem

• Overview - A TL;DR of what happened
• Impact - Data on impact to customers and company
• Root Cause - What caused the incident
• Timeline showing what happened, why it happened, and when it happened
• Details and followups for each incident phase

Incident phases (largely pulled from Meta's post mortem process):

• Detection - Knowing smth bad is happening
• Escalation - Escalating to the appropriate parties to solve it
• Mitigation - Stopping the bleeding
• Remediation - Fixing the wound - anything that got hurt from this
• Prevention - Preventing this class of error from happening again

Todos

• Potentially a Post Mortem template?
• Useful frames like chain of events and 5 whys?