Operations

System health

Data unavailable

Whether each connector is actually reporting, how fresh its data is, and what failed recently. A source that stops updating shows here rather than quietly going stale behind a normal-looking page.

Overall status

Degraded

Connectors reporting

4/5

Database

sqlite

Records held

273,533

Connector status

Live data

Source	Kind	Status	Last successful fetch	Expected cadence
NHC	cyclone	Reporting	2026-10-27 16:03:22 UTC3 min ago	Full advisories every 6 hours, intermediate advisories every 2-3 hours when watches or warnings are in effect
EMSC	earthquake	Reporting	2026-10-27 16:03:21 UTC4 min ago	Near real time; rapid automatic solutions are revised within hours
GEOFON	earthquake	Reporting	2026-10-27 15:58:16 UTC9 min ago	Rapid automatic solutions within minutes; manual revisions follow
USGS	earthquake	Reporting	2026-10-27 16:03:19 UTC4 min ago	Continuous; feeds refresh about every minute
GEM-FAULTS	fault	Stale	never	Released periodically; not a live feed

A connector is marked stale after 6 hours without a successful fetch. Machine-readable status is available at /api/health.

Recent failures

Data unavailable

USGSlive · 2026-10-27 15:52:49 UTCrun failed

HTTP 504 for https://earthquake.usgs.gov/fdsnws/event/1/query?format=geojson&starttime=2026-10-24T15%3A52%3A49.957Z&orderby=time&limit=20000

USGSlive · 2026-10-27 15:47:28 UTCrun failed

HTTP 504 for https://earthquake.usgs.gov/fdsnws/event/1/query?format=geojson&starttime=2026-10-24T15%3A47%3A28.399Z&orderby=time&limit=20000

USGSlive · 2026-10-27 15:41:34 UTCrun failed

HTTP 504 for https://earthquake.usgs.gov/fdsnws/event/1/query?format=geojson&starttime=2026-10-24T15%3A41%3A34.267Z&orderby=time&limit=20000

Source: Prakruti ingest_run log · failed windows are retried on the next scheduled run and are never silently skipped