Metrics and health

Metrics and health

Health check

GET /healthz always returns 200 OK, regardless of --url-path, and never needs credentials. Use it for container, load-balancer and Kubernetes probes.

Prometheus metrics

Off by default. Enable with --metrics; the endpoint is <url-path>/metrics, behind Basic Auth when -U/-P are set.

Metric	Type	Meaning
frontail_build_info{version}	gauge	always 1
frontail_uptime_seconds	gauge	seconds since start
frontail_connected_clients	gauge	browsers on the log socket
frontail_active_reads	gauge	full-log reads in progress ("Start" in the UI); capped at 8 at once
frontail_sources	gauge	files, containers, journal and ssh sources
frontail_lines_total{source}	counter	lines read, per source
frontail_tail_errors_total	counter	errors from tail, the container engine, journalctl or ssh

scrape_configs:
  - job_name: frontail
    static_configs:
      - targets: ['frontail:9001']
    # basic_auth:
    #   username: admin
    #   password_file: /etc/prometheus/frontail_password

Useful alerts: rate(frontail_tail_errors_total[5m]) > 0 (a source is failing), or rate(frontail_lines_total[5m]) == 0 for a source that should always be busy.