Operations
On-call rhythm
- Triage first, then narrow the blast radius.
- Capture the incident timeline in the first 15 minutes.
- Keep customer-facing updates short and honest.
Post-incident
- Document what failed.
- Capture the fix and rollback path.
- List follow-up actions with owners.
Useful habits
- Use a single source of truth for dashboards.
- Keep runbooks executable, not theoretical.
- Prefer simple, observable systems over clever ones.