The Art of Event Driven Observability with OpenTelemetry
summary
Shows concrete instrumented code using the OpenTelemetry SDK with batch span processor, exporter and explicit sampling (trace-id ratios like 20%) to publish trace and span IDs as message properties into a Solace Pub/Sub demo; an initial end-to-end trace ties publisher and consumers into a single 4‑minute span that hides publisher latency, so a second implementation uses span links to create three linked traces (publisher + two consumers) which surfaces per-component timings but loses easy insight into how many consumers a single message triggered and complicates sampling.
AI-written summary · about
chapters (30)
- 0:00intro
- 0:21preface
- 0:45about henrik
- 1:22agenda
- 2:03once upon a time
- 2:2724 years ago
- 3:1520 years ago
- 3:4413 years ago
- 4:2010 years ago
- 4:53what i wanted to represent
- 5:46observability pillars
- 6:36why do we need several signals?
- 7:12we isolate our signals
- 7:55opentelemetry
- 8:43the main component of open telemetry
- 9:28opentelemetry the standard for observability
- 10:04how to produce traces?
- 11:12what is a trace?
- 12:07tracing instrumentation
- 13:04resource
- 13:27tracing sampler
- 15:19what is propagation?
- 16:06distributed tracing in normal architecture
- 17:00distributed tracing in eda
- 17:39example 1 - end2end trace - demo
- 21:19is generated distributed traces useful?
- 21:29example 1 - the usage of span links - demo
- 23:54conclusion, pros/cons
- 25:06take away
- 25:38is it observable?
related talks
41mTracing event-based systems with OpenTelemetry
24mObservability and instrumentation via OpenTelemetry
23mEnhancing Asynchronous Communication Observability with OpenTelemetry
27mAchieving End-to-end Visibility into Complex Event-sourcing Transactions using Distributed Tracing
30mBuilding Observable Streaming Systems with OpenTelemetry
01h 28mUnderstanding Distributed Systems with OpenTelemetry
38mPractical introduction to OpenTelemetry tracing
50mGuide to Observability with OpenTelemetry