Incident Log
An append-only board of cockpit incidents, ranked by severity and grouped by lifecycle state. Each card carries a short ordered timeline — open, mitigate, resolve — and a duration measured against a fixed reference time so the view renders the same every load. This is illustrative project-tracking state, not product data.
By Severity
How the log breaks down across severity tiers. Bars are scaled against the busiest tier.
By State
Incidents grouped by lifecycle — active work first, resolved last. Timelines read top-to-bottom in the order events were logged.
Heartbeat stream stalled for room cohort B
- Area
- Realtime / presence
- Owner
- On-call commander
- Open for
- 1h 0m
- Opened: presence heartbeats flatlined.
- Paged on-call; readiness ceiling pinned.
- Root cause suspected: stale socket pool.
Validation engine rejected a batch of valid uploads
- Area
- Validation engine
- Owner
- Validation lead
- Open for
- 4h 0m
- Opened: false-negative spike on check #4.
- Rolled back rule revision; backlog reprocessing.
- Mitigated: error rate back under budget.
Signature UI flashed a stale error contract
- Area
- Web / signature
- Owner
- Web on-call
- Lasted
- 1h 0m
- Opened: cached error copy after retry.
- Cache key corrected behind flag.
- Resolved: verified on staging + prod.
Cost dashboard double-counted worker minutes
- Area
- Observability / cost
- Owner
- Platform on-call
- Lasted
- 2h 0m
- Opened: cost tile drifted 2x from ledger.
- Found duplicate emit on retry path.
- Resolved: dedupe shipped, tile reconciled.
Incident data is illustrative cockpit project-tracking state authored in src/data/incidents, not product data. Durations are measured against a fixed reference time, so the board is fully deterministic. See the escalation policy for how stalls climb into incidents.