HAProxy integration

HAProxy sits in front of your applications and distributes traffic. If it’s unhealthy, everything behind it is affected. The integration gives you visibility at every layer.

Frontend metrics show what’s coming in: connection counts, request rates, bytes transferred. Backend metrics aggregate across your server pools: how many servers are active, what are queue depths, how fast are backends responding. Server-level metrics drill into individual backends: up or down, connection counts, response characteristics.

Session tracking shows concurrency, revealing how many simultaneous connections you’re handling and what your capacity headroom looks like. Health check metrics surface backends that are failing health checks, often before users notice problems.

Setup requires enabling HAProxy’s stats interface, either the stats socket or the HTTP stats endpoint, which the integration scrapes for metrics.

For high-traffic sites, this visibility is essential for capacity planning. Queue depths tell you when you’re approaching limits. Backend response times help identify which services are slowing down the whole system. The load balancer layer often gets overlooked in monitoring. This integration makes sure it doesn’t.