Maintenance

Runbooks for planned, recurring operational tasks — the routine work that keeps systems healthy: rotations, upgrades, backups/restores, capacity changes, and scheduled cleanups.

Unlike incidents, these are done on a schedule or on demand, not under pressure. A good maintenance runbook states:

• When it runs (cadence or trigger) and who owns it.
• Pre-checks — what must be true before starting.
• Steps — the exact, ordered, copy-pasteable procedure.
• Verification & rollback — how to confirm success and back out safely.

Available runbooks

No Library-native maintenance runbooks yet. Service-specific maintenance (key rotation, reindexing, scaling) lives with the owning repo — see the cross-repo Runbooks index. Add cross-cutting procedures here.