Integration · Web & proxy

HAProxy

Backend health, connection rate, queue depth.

What gets collected

The stats socket or HTTP stats endpoint. Per-backend/per-server state (UP/DOWN/MAINT/DRAIN).

Requirements

  • HAProxy 2.x
  • stats socket or HTTP stats
  • listen block names

Key metrics

Collected out of the box. Every metric is available in dashboards, alerts and PromQL.

haproxy.backend.uphaproxy.session.ratehaproxy.queue.depthhaproxy.response.timehaproxy.5xx.rate

Setup steps

  1. 1stats socket /var/run/haproxy.sock
  2. 2stats listener :9090
  3. 3Frontend label from the listen-block names

After install

  • · Metrics show up on the dashboard within 15-30 seconds
  • · Integration labels flow into filters automatically
  • · Dashboard presets for common metrics are available in the editor
More in the docs
HAProxy
Open the guide →

Frequently asked questions

What does Unimoni collect from HAProxy?
Backend health, connection rate, queue depth. Every metric lands in the same TSDB as the rest of your stack, so you can build dashboards and alerts across HAProxy and everything else in one place.
How long does it take to connect HAProxy?
Usually a few minutes: a handful of steps, no code changes. Metrics show up on your dashboards within 15–30 seconds of the first scrape.
What access does the HAProxy integration need?
Only the minimum listed under Requirements above — a scoped token or role, nothing more. Unimoni asks for the least privilege each system supports.
Is HAProxy monitoring available self-hosted and on the free plan?
Yes. Every integration works identically on Unimoni Cloud and self-hosted, and is included from the free tier up.

Similar integrations