Notifications

Crumb can notify you when something happens, a motion event, or something going wrong with the system itself, through the admin console's Notifications panel.

Channels

A channel is a destination: ntfy, Pushover, a generic webhook, or one of several chat integrations. Add a channel, then send a test notification to confirm it delivers before relying on it.

Rules

Per-camera rules control which cameras notify, and when, including quiet hours so a camera that's fine to notify during the day doesn't page you overnight. Rules are what keeps notifications useful rather than overwhelming: motion detection alone is noisy (wind, passing cars, shadow movement), so sensible defaults and per-camera tuning both matter here as much as they do for the underlying detector, see Tuning.

System alerts

Separately from per-camera motion notifications, a set of system-health alerts is mostly on by default and rule-based: a recorder that's stopped heartbeating, a camera that's stopped writing new footage, low disk space, a stale or failed database backup, or a disconnected object-detection integration. These fire to your configured channels the same way motion notifications do.

One limitation worth knowing: the alerting engine runs inside the API process itself, so it can report problems with the recorder, a camera, or disk space, but it cannot report the API process itself being down. If you want coverage for that case too, a small external uptime check hitting the server's health endpoint from a different machine closes that gap; see docs/AI-INSTALL.md in the repository for the exact endpoint and setup.

Maintenance windows

Before a deliberate restart or planned maintenance, arming a maintenance window suppresses health alerts (they're still recorded, just not dispatched) for the duration, so a normal, expected restart doesn't page anyone for the brief gap while streams reconnect.