How to Monitor HAProxy Load Balancer

Overview

HAProxy is a widely used open-source load balancer and reverse proxy for TCP and HTTP applications. Xitoring's HAProxy integration collects traffic statistics, session counts, error rates, and backend health metrics — giving you full visibility into your load balancing infrastructure.

What Can It Monitor?

Queued Requests — Requests waiting in the queue for a backend server
Sessions — Active sessions across frontends and backends
Request Bytes — Incoming data throughput
Response Bytes — Outgoing data throughput
Requests per Second — Rate of incoming requests
2XX Requests — Successful response count
4XX Requests — Client error response count
5XX Requests — Server error response count
Last Sessions — Most recent session activity
Rate — Current connection rate

Prerequisites

You need to enable HAProxy's built-in statistics page before activating the integration.

On Linux

Edit your HAProxy configuration file:

vim /etc/haproxy/haproxy.cfg

Add the following stats configuration:

listen stats
    bind localhost:8404
    stats enable
    stats uri /

Restart HAProxy:

sudo systemctl restart haproxy

The statistics page is now available at http://127.0.0.1:8404/?stats.

On Windows

Open your HAProxy configuration file (haproxy.cfg) in a text editor and add the same stats configuration:

listen stats
    bind localhost:8404
    stats enable
    stats uri /

Restart HAProxy for changes to take effect.

How to Activate the Integration

Run the Xitogent CLI:

xitogent integrate

Select HAProxy from the list of available integrations. When prompted, enter the URL where the statistics page is served:

http://127.0.0.1:8404

The config file is created at /etc/xitogent/integrations/haproxy_integration.conf.

Setting Up Triggers

Available trigger parameters:

Queued Requests
Sessions
Request Bytes / Response Bytes
Requests per Second
2XX / 4XX / 5XX Request counts
Last Sessions
Rate

Navigate to Triggers on your server page, select HAProxy, choose a metric, set your threshold, and configure notification channels.

Tips

Always bind the stats page to localhost to prevent public exposure
Monitor 5XX Requests to catch backend failures immediately
Set alerts on Queued Requests — a growing queue means backends can't keep up
Track Sessions to plan capacity and detect traffic spikes
Works on both Linux and Windows servers