Runbook Drift Detector for Stale Remediation Steps

Keeps runbooks honest. On a schedule, the agent reads each Confluence runbook, extracts the concrete resources it tells operators to touch — dashboards, services, commands — and checks them against live signals from Datadog. When a step references something that no longer exists or has been renamed, it files a Linear ticket flagging the stale step and its location.

Use it when responders lose trust in runbooks because steps reference dead dashboards or retired services. Best for ops teams that want documentation maintained proactively instead of discovered broken mid-incident.

1. 1A scheduled trigger fires the audit (e.g. weekly).
2. 2The agent lists and reads runbooks from Confluence, capturing each step and version.
3. 3It cross-checks referenced Datadog monitors and dashboards for existence.
4. 4A logic step isolates steps whose referenced resources are missing or renamed.
5. 5For each drift, it opens a Linear ticket naming the runbook, step number, and the broken reference.