All terms
Glossary
Runbook
A short instruction for a recurring operational case: what to check, what to do, whom to call.
A runbook is not system documentation, it is a set of actions for a concrete case: disk full, queue backing up, certificate expired.
Its value shows at three in the morning, when the person on the phone is not the one who built the system.
How you notice it
- The same person is called for every incident.
- Recurring cases are thought through from scratch each time.
- On call duty should be shared across more people.
Frequently asked
How detailed does a runbook need to be?
Short enough to be usable at three in the morning. One page per case: symptom, first check, action, who to call. Anything longer does not get read, anything shorter does not help.