OpenTelemetry Tracing — Real Challenges in Complex Systems
summary
The talk examines real-world challenges when deploying OpenTelemetry tracing in complex architectures, explaining how tracing works under the hood and why traces get lost or broken. It covers concrete fixes for multithreading, Outbox/Inbox patterns (manual and with Debezium), Kafka Streams and batch processing, and sagas; plus strategies to control data volume and cost via head- vs tail-based sampling, OpenTelemetry Collector configuration, and filtering, with production code, configs, pitfalls and effective tactics.
AI-written summary · about
chapters (12)
- 0:00OpenTelemetry Tracing intro
- 1:10Distributed tracing basics
- 9:37Tracing challenges
- 12:06Tracing standardization history
- 20:35Tracing in multithreaded apps
- 27:58Tracing with Outbox Pattern
- 37:12Span Links advanced
- 39:39Sampling trace data
- 44:37Collector role in OTel
- 50:14Tracing benefits
- 53:01Tracing implementation tips
- 55:07Q&A
related talks
50mOpenTelemetry Tracing — It's Not That Simple
52mHow to effectively spy on your systems
35mOpenTelemetry for Busy People
38mPractical introduction to OpenTelemetry tracing
43mPractical Introduction to OpenTelemetry Tracing
01h 28mUnderstanding Distributed Systems with OpenTelemetry
33mPractical introduction to OpenTelemetry tracing
34mReduce Complexity with Distributing Tracing