APM and tracing basics

Topic: Monitoring basics

Application Performance Monitoring and distributed tracing show request flow and latency across services. Use when you need to find slow or failed requests across a distributed system.

Intent: How-to

Quick answer

  • Instrument app with tracing SDK or agent. Emit spans with trace ID and parent. Send to Jaeger, Tempo, or vendor.
  • Trace one request across services. Find slow span or error. Use for debugging and SLO.
  • Sample if volume high. Keep sampling rate for errors high. Secure and limit retention.

Prerequisites

Steps

  1. Instrument

    Add tracing SDK or agent. Propagate trace context. Send spans to backend.

  2. Query and debug

    Find trace by ID or filter. Inspect spans and latency. Correlate with logs.

  3. Sample and retain

    Set sample rate. Higher for errors. Set retention. Secure.

Summary

Instrument with tracing; query traces; sample and retain. Use for latency and errors.

Prerequisites

Steps

Step 1: Instrument

Add SDK or agent. Propagate context. Send spans.

Step 2: Query and debug

Find trace. Inspect spans. Correlate with logs.

Step 3: Sample and retain

Sample rate. Retention. Secure.

Verification

  • Traces visible. Can find slow or failed request.

Troubleshooting

No traces — Check instrumentation and backend. Too much data — Lower sample rate.

Next steps

Continue to