Platform status
Component health plus latency measured from real API telemetry (last 24h, 5-minute refresh). Formal SLA numbers publish after a sustained external-probe window — until then this page is the honest live view.
REST API v1OPERATIONAL
MCP serverOPERATIONAL
Data warehouseOPERATIONAL
Data freshnessDEGRADED
Measured latency (last 24h, 1 calls)
| Endpoint | Calls | p50 | p95 | 5xx rate |
|---|---|---|---|---|
| /api/v1/area | 1 | 8848ms | 8848ms | 0% |
Data pipeline freshness
14 of 15 monitored dataset families are current; 1 stale, 0 failed and 0 awaiting a registered successful run. Per-dataset row counts, validation state, ingestion timestamps and refresh cadences live on the data catalog, and every API response carries them machine-readably in meta.vintages.
Probes: /api/health/v1 (component health, unauthenticated) · /api/v1/area with a monitoring key (the real sold surface). Incidents will be noted here and in the changelog.