Guide
Create Your First Runbook
Build a runbook that is specific enough to help during a real incident and simple enough to trust.
Before you start
- Pick one repeated workflow you actually run during incidents.
- Decide which systems, functions, environments, tags, or teams the runbook should target.
Do this
- Open
Runbooksand create a new runbook. - Name the runbook after the job it does, not the team that wrote it.
- Add a short description so responders know when to use it.
- Set the target systems, functions, environments, tags, or teams the runbook should match.
- Add the steps in the order responders expect them to happen.
- Keep the first version short and focused on one operational outcome.
- Save the runbook and leave it inactive until the steps and dependencies look right.
- Activate it only when the runbook is ready to launch during a live incident.
Check it worked
- The runbook appears in the library with the expected targets.
- Each step shows as ready instead of blocked.
- The runbook is easy to scan and does not ask responders to do unrelated work.
If it does not work
- If the runbook is too long, split it into smaller runbooks by job.
- If steps are blocked, fix the integration or replace the step before you activate the runbook.
- If responders would not know when to use it, rewrite the name and description first.
