Stack Status¶
What this is¶
Admin health check for the PiHerder stack itself (web, Postgres, Redis, Celery, scheduler, disk free) — not remote fleet SSH probes.
Why it exists¶
When backups stay pending or the UI feels dead, you need a single place that answers “is Redis up? is Celery registered? is /backups full?” before you dig into host troubleshooting.
Where: Settings → Status (admin).
What is checked¶
| Component | Idea |
|---|---|
| Web | Process / health |
| PostgreSQL | SELECT 1 |
| Redis | Broker ping |
| Celery | Nodes + pool slots (CELERY_CONCURRENCY) |
| APScheduler | Running + jobs registered |
| Disk (fast) | Mount free on /backups, /data, /herder_backups |
| Disk (lazy) | Full tree du + host folders as a dense list — View details only |
- Manual Check now + ~2 minute scheduled poll.
- Notifications on state transition only (no spam).
- Metrics gauges can reflect last check.
Host dependency check (fleet)¶
Separate from stack Status: per-server remote tools for enabled features.
Server detail → Re-check · also after SSH test / key deploy / least-priv.