Status
The price archive is the only thing here that cannot be rebuilt. Neither venue publishes historic depth, so time the collector spends stopped is permanently missing rather than merely delayed, and a stopped collector is silent, because every page carries on serving live venue data exactly as before.
no successful collection for 41 minutes — the archive is gaining a permanent gap
Recent runs
| Started | Scope | Polled | Written | Errors | Result |
|---|---|---|---|---|---|
| 2026-08-29 04:30:01Z | core | 95 | 47 | 0 | ok |
| 2026-08-29 04:25:01Z | core | 95 | 46 | 0 | ok |
| 2026-08-29 04:20:01Z | core | 95 | 37 | 0 | ok |
| 2026-08-29 04:15:01Z | core | 95 | 61 | 0 | ok |
| 2026-08-29 04:10:00Z | core | 95 | 43 | 0 | ok |
| 2026-08-29 04:05:00Z | core | 95 | 66 | 0 | ok |
| 2026-08-29 04:03:00Z | wide | 312 | 227 | 1 | failed |
| 2026-08-29 04:00:00Z | core | 95 | 44 | 0 | ok |
| 2026-08-29 03:55:01Z | core | 95 | 42 | 0 | ok |
| 2026-08-29 03:50:01Z | core | 95 | 39 | 0 | ok |
| 2026-08-29 03:45:01Z | core | 95 | 50 | 0 | ok |
| 2026-08-29 03:40:00Z | core | 95 | 38 | 0 | ok |
Watching this from outside
/api/health/collection returns 503 once collection has stalled and 200 otherwise. Point any uptime service at it and the alert means something specific.
It is deliberately separate from /api/health, which answers whether the site is up, and stays true while collection is dead. A monitor pointed at that would report green through a week-long hole in the archive, which is precisely the failure this is meant to catch.