APM (Application Performance Monitoring)

The APM module provides application-level performance monitoring by combining metrics, traces, and logs into a unified service health view. It automatically discovers services, maps dependencies, and tracks SLIs like latency, error rate, and throughput.

Key Features

• Auto-discovery — Automatically detects services and endpoints from incoming trace and metric data
• Service overview — Health dashboard per service showing latency, error rate, throughput, and Apdex score
• Endpoint performance — Drill into individual endpoints to see latency distributions and error rates
• Dependency mapping — Auto-generated service dependency graph based on observed trace data
• SLO tracking — Define SLOs per service and track burn rate with automated alerting
• Deployment tracking — Correlate performance changes with deployment events
• Database query analysis — Surface slow database queries captured in trace spans
• Runtime metrics — JVM, Node.js, Python, Go, and .NET runtime metrics (GC, memory, threads)

How to Access

Navigate to APM in the left sidebar. The overview page shows all discovered services with health indicators.

Basic Usage

1. Open APM from the sidebar to see the service catalog
2. Click a service to view its health dashboard (latency, errors, throughput)
3. Navigate to the Endpoints tab to see per-endpoint performance
4. Click an endpoint to view trace samples and latency distributions
5. Go to the Dependencies tab to see upstream and downstream services
6. Set up SLOs under the service’s SLO tab (e.g., “p99 latency < 200ms for 99.9% of requests”)
7. Enable Deployment Markers under Settings → Integrations to correlate releases with performance