How To Troubleshoot and Monitor Applications using OpenTelemetry
summary
Shows exactly what gets built and what it actually produces: a local demo (Docker Compose) that runs Jaeger, Zipkin, Prometheus and an OpenTelemetry Collector, instruments a Flask app three ways—manual instrumentation, contrib instrumentation, no-code instrumentation—and exports OTLP traces and Prometheus metrics. Manual yields verbose boilerplate and thin spans; contrib captures headers, status and richer attributes; no‑code wraps apps without source edits. Collector buffers and forwards traces; the demos successfully show errors, trace IDs, parent/child spans and basic sampling behavior.
AI-written summary · about
chapters (21)
- 0:00Introduction to observability
- 1:20The three pillars of MELT
- 6:01OpenTelemetry high-level view
- 8:51Technical architecture details
- 13:35OpenTelemetry collector
- 16:26Setting up the demo environment
- 19:16Docker compose configuration
- 24:10Configuring the collector
- 28:15Starting the backend services
- 30:18Base case: flask application
- 33:34Testing the uninstrumented app
- 41:48Manual instrumentation
- 48:19Testing manual instrumentation
- 56:35Contrib instrumentation
- 1:02:03Analyzing contrib data
- 1:12:33No code instrumentation
- 1:27:46Testing no code setup
- 1:32:10Advanced instrumentation features
- 1:40:55Distributed tracing in action
- 1:45:44Security and configuration
- 1:54:06Conclusion and Q&A
related talks
23mModern Application Debugging: An Intro to OpenTelemetry
35mOpenTelemetry for Busy People
38mPractical introduction to OpenTelemetry tracing
22mOpenTelemetry – An Observability Framework for Cloud-Native Software
42mObservable Applications Using OpenTelemetry
34mOpenTelemetry for Dummies: Instrumenting Go Apps
41mModern Application Debugging: An Introduction To OpenTelemetry
50mGuide to Observability with OpenTelemetry