Observability For (Java) Devs - 2023 Edition
summary
Argues that observability must be instrumented out of the app and sent to a separate system, shows a working Docker Compose demo with an openTelemetry collector and Jaeger/Tempo back end, and reports adoption metrics (New Relic sees a 30x rise in collectors and one third of its traffic in 12 months; Gartner predicts ~50% of observability traffic). Recommends Micrometer for Java metrics, warns against single-core containers for modern GCs, describes a production-ready Java agent auto-instrumentation and OTLP as OTLP over HTTP/2 (protobuf) with JFR integration moving from contrib into mainline.
AI-written summary · about
chapters (14)
- 0:00Introduction and speaker background
- 3:10What is observability?
- 10:35Why observability matters now
- 16:32The three pillars of observability
- 23:09Metrics in Java with Micrometer
- 25:33Logs and their role in observability
- 27:20Distributed tracing explained
- 37:13Actionable insights from observability data
- 41:04Anti-patterns to avoid
- 45:20OpenTelemetry overview and market shift
- 50:37OpenTelemetry components and Java support
- 57:18OpenTelemetry protocol and Java performance
- 1:00:51Future of observability and OpenTelemetry
- 1:02:20Q&A session
related talks
51mFull-Stack Observability: One Tool to Rule Them All
48mAPI Observability by Design: Building Systems That Tell Their Own Stories
58mThrough the looking glass: Effective observability for cloud native applications
38mObservability 101 with Spring and Micrometer
47mThrough the Looking Glass: Effective Observability for Cloud-Native Applications
38mBecoming a cloud-native doctor
37m5 reasons why (java) developers don’t buy in the observability hype
26mMake Monitoring Easy for your Java Apps