Monitoring

Health endpoints

Endpoint	Meaning
GET /api/health	Unauthenticated; {"ok":true} when the API is up
Plugin host GET /health	Unauthenticated, reachable only on the internal network: {ok, version, sandbox}

Docker Compose runs a health check on the plugin host; docker compose ps shows it.

Prometheus

GET /api/metrics returns Prometheus text. It is administrator only and, today, only an administrator session can read it: API tokens cannot reach it. That makes unattended scraping awkward (a scraper has to hold an administrator session cookie). Until a dedicated metrics token exists, either scrape from a trusted small script that reuses a long-lived administrator session, or use the notifications and the Resilience page for alerting. The metrics:

Metric	Labels	Meaning
fledge_nodes	status	Registered nodes by connection state
fledge_servers	status	Servers by observed state
fledge_jobs	state	Jobs by state
fledge_failover_events	state	Failover events by state
fledge_servers_without_backup		Servers that have no successful backup
fledge_servers_crash_looping		Servers whose automatic restarts were stopped
fledge_notification_channels_failing		Channels whose last delivery failed
fledge_schedule_runs	state	Schedule runs in the last 24 hours
fledge_plugins	enabled	Installed plugins
fledge_server_addons	state	Tracked add-ons
fledge_sessions, fledge_passkeys		Active sessions and registered passkeys

Example alert rules:

- alert: FledgeNodeDown
  expr: fledge_nodes{status="disconnected"} > 0
  for: 5m
- alert: FledgeServersWithoutBackup
  expr: fledge_servers_without_backup > 0
  for: 1d
- alert: FledgeCrashLoop
  expr: fledge_servers_crash_looping > 0

History in the panel

Each server's Usage tab keeps CPU and memory history (1 hour to 30 days; see Usage history).

Notifications

Instead of polling, let Fledge tell you: Notifications cover node offline/online, crashes, crash loops, low disk, failed backups, failover, failed schedules and plugin problems.

Logs

docker compose logs -f api web plugins
docker compose logs --since 10m api

API logs are JSON lines. On a node: journalctl -u fledge-agent -f. A plugin's own log lines appear in Plugins → (plugin) → Logs.