Distributed Tracing in Action: Our Journey with OpenTelemetry
summary
ThousandEyes describes a concrete pipeline: Zipkin from Istio/EngineX plus OTLP from apps flow through an OpenTelemetry Collector (Kubernetes operator) into Grafana Tempo, with B3 multi and W3C propagators to keep traces connected. An injected Java agent auto-instruments apps so over 100+ services are instrumented quickly, staging samples 100% and prod ~1% with plans for tail sampling. Real wins include pinpointing a UI-caused 500 by following a trace to an exception and trimming Jenkins builds by 1–2 minutes after tracing a slow file read.
AI-written summary · about
chapters (19)
- 0:00Distributed tracing journey
- 1:22ThousandEyes platform overview
- 4:54Troubleshooting without tracing
- 7:43How tracing improves troubleshooting
- 9:48Context propagation in tracing
- 11:51Why auto-instrumentation isn't enough
- 13:35OpenTelemetry's role in tracing
- 15:20Setting up OTel collector pipeline
- 16:45Deploying collectors with K8s operator
- 17:49Configuring Java auto-instrumentation
- 19:13Connected trace architecture
- 20:36Instrumentation adoption strategies
- 22:04Encouraging team instrumentation
- 23:30Driving usage through documentation and demos
- 24:54Mixing signals for trace discovery
- 25:57Training responders with simulations
- 27:20Future tracing and OpenTelemetry plans
- 29:45Recap and key takeaways
- 31:10Closing and resources
related talks
08mLightning Talk: Kick-Start Distributed Tracing in Istio with OpenTelemetry
49mDistributed Tracing in Java? OpenTelemetry for the Rescue!
01h 28mUnderstanding Distributed Systems with OpenTelemetry
47mUsing metric and traces to diagnose our cloud-native applications
22mOpen Telemetry for API Monitoring
59mOpenTelemetry will save your day and night
47mI'm Tracing You!
37mThrough the looking glass: Effective observability for cloud native applications