Live health of every adapter + poller in the ingestion pipeline. Auto-refreshes every 5s.
One row per upstream worker. Parsed metrics come from the orchestrator’s heartbeat message.
No worker heartbeats in the database yet.
Start an ingestion worker with pnpm ingest — rows will appear here within ten seconds.
Row counts across every time-series table. “Last hour” uses ts (ms-unix) or date (today).
Error tracking TBD.
This section will surface rows from ingest_errors once the table exists. In the meantime, check your ingestion logs directly.