Would You Like Some Tracing With Your Monitoring?
summary
Jaeger (OpenTracing-native) is a distributed tracing system used to instrument hundreds of microservices and reveal where transactions actually spend time. Live demo shows a Hot Rod app, dependency graphs, and a trace where a MySQL span takes ~50% of latency, a thread pool of three creates a staircase bottleneck, and baggage: customer ID propagated enables business-level metrics. Lessons: use OpenTracing, enable tracing by default, measure trace quality, and automate black‑box tracing.
AI-written summary · about
chapters (17)
- 0:00Intro
- 3:33How Distributed Tracing Works
- 4:45Demo
- 10:28Performance issues
- 18:00Summary
- 23:11Zero trace
- 23:44Service meshes
- 25:04What we learnt
- 25:38What you should do
- 26:49Distributed context propagation
- 27:23What we use it for
- 29:07Adoption trace quality
- 29:27Trace quality metrics
- 30:49Getting used to tracing
- 31:07Service dependence diagram
- 31:53Sample dependency diagram
- 33:15Outro
related talks
37mOpenTelemetry & Distributed Tracing: A Jaeger v2 Deep Dive
23mIntroduction to OpenTracing
40mCracking Microservices Performance Issues with Distributed Tracing
43mDebugging and Profiling Distributed Applications Using OpenTracing
51mNavigating Distributed Traces
41mPractical introduction to OpenTelemetry tracing
49mDistributed Tracing in Java? OpenTelemetry for the Rescue!
47mUsing metric and traces to diagnose our cloud-native applications