Overview
Varnish is a powerful reverse caching proxy used as an HTTP accelerator for content-heavy dynamic websites and APIs. Xitoring's Varnish integration provides comprehensive monitoring of cache performance, backend health, thread pools, and session states — ensuring your caching layer delivers maximum performance.
What Can It Monitor?
Cache Metrics
- Cache Hit — Total cache hits
- Cache Hit Rate — Percentage of requests served from cache
- Cache Hit Pass per Second — Requests passed through to backend despite cache
- Cache Miss per Second — Requests not found in cache
Backend Metrics
- Backend Connection per Second — New connections to backend servers
- Backend Request per Second — Requests forwarded to backends
- Backend Busy per Second — Backend connection pool exhaustion events
- Backend Fail per Second — Failed backend connections
- Backend Unhealthy per Second — Unhealthy backend detections
- Backend Recycle / Reuse per Second — Connection recycling rates
- Backend Too Late per Second — Expired backend connections
Resource Metrics
- CPU — Varnish process CPU utilization
- Memory — RAM consumed by Varnish
- N Expired per Second — Objects expired from cache
- N LRU Nuked per Second — Objects evicted via LRU (Least Recently Used)
Session & Thread Metrics
- Client Request per Second — Rate of incoming client requests
- Session Connection per Second — New client sessions
- Session Dropped per Second — Dropped client sessions
- Threads per Second — Worker thread creation rate
- Thread Queue Length per Second — Requests queued waiting for threads
- Threads Created / Failed / Limited per Second — Thread lifecycle events
ESI Metrics
- ESI Errors per Second — Edge Side Include processing errors
- ESI Warnings per Second — ESI processing warnings
Prerequisites
None! There are no software prerequisites to enable this integration.
How to Activate the Integration
Run the Xitogent CLI:
xitogent integrate
Select Varnish from the list. Xitogent tests the connection and sets up the rest automatically. Within a minute, graphs and data appear on your server page.
Setting Up Triggers
All metrics listed above are available as trigger parameters. Key ones to set up:
- Cache Hit Rate — Alert when it drops below your target (e.g., 90%)
- Backend Fail per Second — Catch backend connection failures
- Session Dropped per Second — Detect client connection issues
- Thread Queue Length — Alert when requests are queuing for threads
- N LRU Nuked per Second — Detect cache memory pressure
Navigate to Triggers on your server page, select Varnish, choose a metric, set your threshold, and configure notification channels.
Tips
- Varnish integration requires zero configuration — just select it from the CLI
- Monitor Cache Hit Rate as your primary KPI — higher is better
- Set alerts on Backend Fail and Backend Unhealthy to catch origin server issues
- Watch Thread Queue Length — a growing queue means you need more worker threads
- Track N LRU Nuked per Second — frequent evictions mean your cache needs more memory
- Use Session Dropped alerts to detect when Varnish is overwhelmed