Antipatterns in Observability: Lessons Learned and How OpenTelemetry Solves Them
Talkat KubeCon + CloudNativeCon China 2025 · on Tuesday, 2025-06-10· at 13:45· Level 16 | Grand Ballroom I·
summary
Lists concrete anti patterns — incomplete instrumentation that produces disconnected traces, overinstrumentation that explodes cost and query latency, inconsistent metadata names that break correlation, and vendor lockin/tool sprawl that blocks migration — and shows concrete fixes: Java agent runtime injection for zero‑config traces, the open telemetry collector with receivers/processors/exporters to drop spans or filter metrics, and semantic conventions to infer visibility (example: 100% errors from one pod isolated by resource attributes).
AI-written summary · about
chapters (9)
- 0:00Introduction
- 1:01Understanding anti-patterns
- 3:24Impact of anti-patterns
- 7:29Telemetry data anti-patterns
- 14:16Optimizing data collection
- 19:51Naming and standardization
- 23:03Observability platform pitfalls
- 26:46Scalability and culture
- 28:54Conclusion and Q&A
related talks
36mObservability For You and Me with OpenTelemetry
26mYou Can’t Fix What You Can’t See: Tales from the Observability Frontier
02h 41mThe Science of Signals: Mastering Telemetry for Observability
50mGuide to Observability with OpenTelemetry
53mObservability with OpenTelemetry or how to say goodbye to Vendor Lock-in
20mObservability and Open Telemetry
22mOpenTelemetry – An Observability Framework for Cloud-Native Software
52mHow to effectively spy on your systems