Unauthenticated liveness and status, so an uptime checker needs no credential. GET /healthz returns a capability snapshot (provider readiness, queue depth, row counts); /status.json and /status render a wider operational picture with every probe wrapped independently.
Check server health
Liveness plus the numbers an operator needs during an incident: app and subscription counts, the running commit, whether FCM and APNs are configured, and the send queue's depth. No authentication.
The fcm / apns flags are server-wide (the APNS_P8 / FCM environment
credentials), not per app — per-app platform readiness is on
GET /v1/apps/{app_id}/settings under platforms. If the queue backend is itself down,
queue degrades to {"backend": "unavailable", "error": ...} rather than failing the
health check.
Check server health › Responses
Health flags, app/subscription counts and send-queue stats