The gentle touch of APM - how code tracing works in Node.js
summary
Shows how Elastic APM actually instruments Node.js: a small agent module required at startup hooks into frameworks or monkey‑patches libraries, emits transaction and span JSON, captures headers/body/user context and full JavaScript stack frames with file and line via stackman, groups errors and visualizes a service map, and supports distributed tracing by propagating a trace id. Sampling reduces cost and measured overhead is typically single‑digit CPU impact, so it works in production.
AI-written summary · about
chapters (19)
- 0:00Introduction
- 1:56What is APM and tracing
- 5:28Elastic APM architecture
- 9:19APM agents collect data
- 10:02Exploring APM UI
- 21:40Error handling walkthrough
- 24:07Transaction instrumentation
- 25:32APM agent loading
- 26:14Recording errors
- 31:15Request/response context
- 35:34Stack traces in detail
- 39:46Library vs app code
- 40:30APM in Elastic Stack
- 42:16Learn more about APM
- 43:26Q&A: Performance impact
- 45:34Q&A: Adding APM to apps
- 46:59Q&A: Distributed tracing
- 48:23Q&A: Node.js vs RUM
- 50:10Q&A: Node.js observability
related talks
25mTracing Node. The Open Source Edition.
07mInstrumenting Node.js Internals
22mExamining Observability in Node.js
33mDefeating Node.js Malware through API Tracing
28mTracing Node. The Open Source Edition
44mHow to Measure Node js Applications When You Have No Hands
08mComprehensive Observability via Distributed Tracing on Node.js8
17mLogging, Metrics, and Tracing with Node.js