Status · live probes · refreshed at render
System status.
Real probes, not a redirect.
Every component below is probed at request time against the live API. No cached green badges, no marketing redirect — if something is down, this page says so.
Some systems are degraded.
Operational · probed 2026-07-16 17:21:46 UTC
Components
API Gateway
Liveness probe — process is up and accepting requests.
93ms
Operational
Readiness
DB connectivity + audit-chain writability + killswitch state.
3631ms
Operational
Model Catalog
Full routable model set served from /v1/models.
319ms
Operational
Catalog Spine
Shared catalog between router, marketing, and dashboard.
393ms
Operational
Audit Feed
Public signed routing decisions on the audit log.
318ms
Operational
Audit Log Height
Current event height of the public audit log.
timeout
5002ms
Unknown
Leaderboard
Model decision counts, latency percentiles, and cost stats.
298ms
Operational
How we probe
probe matrix7 endpoints · no-store · 5s timeout
# each component is probed at render time
GET https://api.ainfera.ai/healthz
GET https://api.ainfera.ai/readyz
GET https://api.ainfera.ai/v1/models
GET https://api.ainfera.ai/v1/catalog/public
GET https://api.ainfera.ai/v1/audit/public
GET https://api.ainfera.ai/v1/audit/height
GET https://api.ainfera.ai/v1/stats/public/leaderboard
Probes use cache: no-store so you always see current state. A 200 response = operational. 4xx = degraded. 5xx or timeout = down. This page itself is served by Vercel (marketing surface); if Vercel is down, this page won't load — the API health endpoints above are the source of truth.
Incident history
2026-07-12maintenanceResolved
Status page launched
Independent /status page stood up (RED-8). Previously /status redirected to /routing#proof — auditors had no health surface. Now shows live component probes at render time.