Public incident log — UpSkillZone
What has gone wrong, and how we fixed it.
Every operational incident the platform records, sorted newest first. SEV-1 and SEV-2 postmortems publish within 5 business days; SEV-3 within 10. The register is the same hand-curated set the operator surface reads — when /api/v1/admin/incidents ships, both pages will switch over and the rendering shells stay intact.
Open
2
Status not in {resolved}.
SEV-1 / SEV-2 ever
1
Lifetime count across the platform's history.
Year-to-date
5
Incidents started in the current calendar year.
Postmortem SLA
5 / 10d
SEV-1 and SEV-2 ≤5 business days; SEV-3 ≤10.
Incident register
One row per incident. The id links to the public postmortem.
| ID | Title | Severity | Status | Started | Duration |
|---|---|---|---|---|---|
| INC-2026-005 | Bias audit query log gap | SEV-3 | resolved | Apr 21, 2026, 01:55 PM UTC | 1h 7m |
| INC-2026-004 | JWKS endpoint cold cache | SEV-3 | monitoring | Apr 9, 2026, 05:18 AM UTC | 28m |
| INC-2026-003 | JWT refresh rotation race | SEV-2 | mitigated | Mar 18, 2026, 10:07 PM UTC | 1h 27m |
| INC-2026-002 | Talent search returned stale rankings | SEV-3 | resolved | Feb 3, 2026, 09:41 AM UTC | 1h 34m |
| INC-2026-001 | Mentor queue worker drained slowly | SEV-3 | resolved | Jan 12, 2026, 02:22 PM UTC | 1h 46m |