Observable Applications Using OpenTelemetry
summary
Explains why distributed tracing is worth the cost, shows OpenTelemetry as the merged standard replacing OpenTracing/OpenCensus, and demos instrumenting a Dockerized fake job-title app with Go and Python libraries sending spans to a local Jaeger backend. Shows concrete code: tracer.Start, defer span.End, attributes/events, and using inject/extract for gRPC/HTTP context propagation via interceptors. Covers SDK exporters, sampling to reduce load, bridges for gradual migration, and that auto-instrumentation is in progress.
AI-written summary · about
chapters (24)
- 0:00Intro
- 2:11Distributed Tracing in 30 Seconds
- 3:04Logs, Metrics, Traces
- 4:10What is the Question?
- 5:51It's a lot of Work
- 6:37We Can't Vendor-Lock
- 7:28Distributed Tracing vs. Microservices
- 8:31Multiple Everything
- 10:07Introducing OpenTelemetry
- 11:19OpenTelemetry - Architecture
- 13:51Separation of Concerns
- 14:36Protecting User Applications
- 15:39Project Status
- 16:06Go Library Status
- 16:31Python Library Status
- 16:49Instrumenting Go Code
- 18:01Propagating Context Between Processes
- 18:50Instrumenting Python Code
- 19:26Demo: Fake Job Title Generator
- 28:22Context - Request-Scoped Data
- 29:10In-Process Context Propagation
- 32:23Distributed Context Propagation
- 34:14Conclusion
- 35:47Thank you! Johannes Liebermann
related talks
34mOpenTelemetry for Dummies: Instrumenting Go Apps
01h 28mUnderstanding Distributed Systems with OpenTelemetry
38mPractical introduction to OpenTelemetry tracing
30mMastering Distributed Tracing with Go
02h 02mHow To Troubleshoot and Monitor Applications using OpenTelemetry
41mModern Application Debugging: An Introduction To OpenTelemetry
33mPractical introduction to OpenTelemetry tracing
30mPractical introduction to OpenTelemetry Tracing