Tracing event-based systems with OpenTelemetry
summary
Shows how to get end-to-end traces through Kafka by adding an OpenTelemetry Java agent for zero‑code tracing or by adding SDK calls to create custom spans in a Spring Boot producer/consumer demo running on AWS AppRunner and Confluent Kafka. Exports use the OTLP endpoint (example wired to New Relic). Demo surfaces a ~6.6s consumer lag, uninstrumented hotspots that manual spans reveal, and how to scrape broker metrics or instrument Confluent/Flink; New Relic cites processing 50M messages/sec (~1 TB/s aggregate).
AI-written summary · about
chapters (9)
- 0:00Introduction
- 3:19Tracing event systems
- 7:13OpenTelemetry fundamentals
- 16:00Zero-code vs manual
- 21:37Demo: Kafka tracing
- 27:38Manual instrumentation
- 33:58Custom attributes
- 37:11Monitoring Kafka brokers
- 39:35Summary
related talks
43mTracing Kafka with OpenTelemetry
38mPractical introduction to OpenTelemetry tracing
30mBuilding Observable Streaming Systems with OpenTelemetry
47mI'm Tracing You!
51mPractical introduction to OpenTelemetry tracing
27mAchieving End-to-end Visibility into Complex Event-sourcing Transactions using Distributed Tracing
26mThe Art of Event Driven Observability with OpenTelemetry
49mDistributed Tracing in Java? OpenTelemetry for the Rescue!