Browse
Incident Management & Postmortems
Running an incident from detection to resolution, and writing a blameless postmortem that actually prevents a repeat.
What it is
Incident management is the process of responding to a live outage or degradation; a postmortem is the structured write-up afterward that turns the incident into a durable improvement.
Key points
- Incident commander: a designated role during a major incident who coordinates response and communication, freeing responders to focus on fixing the problem rather than status updates.
- Mitigate first, root-cause later: the immediate priority during an incident is restoring service (rollback, failover, feature flag) — full root cause analysis happens after, not during.
- Blameless postmortems: focus on what in the system or process allowed the failure, not who made a mistake — this is what makes people willing to report and discuss failures honestly instead of hiding them.
- A good postmortem produces concrete, owned, tracked action items — a document that's read once and forgotten hasn't actually prevented anything from recurring.