Writing a runbook
At 3am, under pressure, nobody remembers the exact steps. A runbook is those steps written down while you were calm, for the moment you are not.
What belongs in one
Anything you would hate to improvise: a failover, a restore, bringing the stack back in order, rotating a leaked key. Write the concrete commands and the order, the checks that confirm each step worked, and who to call when it does not. A runbook is for the tired and the stand-in, so leave nothing to "obviously".
Make it usable under pressure
- Numbered steps, one action each, in the real order.
- The exact command, not a description of it.
- A check after each step so you know it worked before moving on.
- A clearly marked point to stop and escalate.
Keep it honest
A runbook rots the moment the system changes under it. Tie it to disaster recovery and actually rehearse it — the drill both proves the steps and keeps them current. A runbook you have never run is a guess in a nicer format.
Still stuck? Mail support@novogara.com — an engineer answers, at any hour. Back to the knowledge base