Metrics

/metrics on the dashboard’s address serves Prometheus text:

• queue depth and the age of the oldest message, per server;
• blocklist listings, per IP and zone;
• delivery outcomes, by server and status;
• Mailinois’s own health: syslog frames received, lines dropped, write errors and the buffer.

The metrics

These are the names in the current version. curl http://localhost:7480/metrics shows exactly what yours exports.

Metric	Type	Labels
mailinois_queue_messages	gauge	server	Messages still in a Postfix queue.
mailinois_queue_oldest_seconds	gauge	server	Age of the oldest message in the queue, 0 when it is empty.
mailinois_dnsbl_listed	gauge	ip, zone	1 when the latest check found the IP on that blocklist.
mailinois_alert_firing	gauge	alert	1 while an alert is firing.
mailinois_outcomes_total	counter	server, status	Delivery attempts and rejections ingested since start.
mailinois_syslog_frames_total	counter		Syslog frames received.
mailinois_syslog_unparsed_total	counter		Frames that were not syslog.
mailinois_lines_dropped_total	counter		Lines and captures dropped, because the buffer was full or the database was away too long.
mailinois_ingest_errors_total	counter		Failed writes to the database.
mailinois_captures_total	counter		Captured messages saved.
mailinois_lines_buffered	gauge		Lines waiting to be written.

Scraping

/metrics sits behind the same authentication as the API. With auth.mode set to off it is open; otherwise give Prometheus a token with the read scope:

prometheus.yml

scrape_configs:
  - job_name: mailinois
    authorization:
      type: Bearer
      credentials_file: /etc/prometheus/mailinois-token
    static_configs:
      - targets: ["mailinois.internal:7480"]

/healthz, on the same address, answers health checks without a login: ok while the database is reachable, and 503 when it is not.