Alerts and monitoring

Alerts for the admin

Telegram → Infrastructure → Private alerts: the “Connect chat” button gives a one-time link, open it in the bot. After that, pick the events:

• a node is unavailable or recovered;
• WARP stopped working or recovered, with the reason;
• a cascade exit node is unavailable or recovered;
• a connection is unavailable or recovered;
• autotune changed a connection or could not restore it;
• a TLS certificate problem;
• an update failed or rolled back;
• a torrent was caught.

For failure events a recovery message arrives too. You can also receive daily database backups here.

Public status in a channel

Public status channel shows subscribers which servers are working.

1. Add the bot to the channel as an administrator with permission to post, edit and pin messages.
2. Enter the channel: a public @username or a numeric ID.
3. Choose “Pinned summary” (one message the bot keeps updating) and/or separate messages about changes.

The status shows only nodes that have Name in status channel filled in their settings. Subscribers see this name, and server addresses are not published.

Prometheus

The panel serves metrics at /api/v1/metrics on the API address, and you need a key with “Read” access. Metrics are split by node and connection, but never by user: the number of users by state, online, traffic, node and connection availability, node load.

The API page (Settings → Security → API) shows a ready block for prometheus.yml, roughly like this:

scrape_configs:
  - job_name: mikan
    scheme: https
    metrics_path: /SECRET-PATH/api/v1/metrics
    authorization:
      credentials: mk_...
    static_configs:
      - targets: ['vpn.example.com:PORT']