Examining Observability in Node.js
summary
Shows concrete Node.js probes: performance_hooks, trace events, heap snapshot and V8 inspector are used to measure latency, memory and CPU; demoed perf marks, trace files and heap dumps. External agents include blocked, New Relic, Datadog, Instana, Dynatrace and App Dynamics; an alternative runtime called solid (nsolid) avoids agent overhead. Reported numbers: vanilla Node 30ms startup, solid ~35–40ms, App Dynamics/DataDog much slower; solid adds ~2MB RAM vs New Relic 15MB and DataDog 57MB; solid reaches ~10k RPS.
AI-written summary · about
chapters (17)
- 0:00Start
- 0:45Agenda
- 4:22Choosing an Observability Tool
- 4:34Performance Hooks
- 7:04Profiling
- 7:32Trace Events
- 8:15Chrome Tracing
- 9:12Heap Snapshot
- 11:03Inspect Break Flag
- 11:23Demo
- 14:03Blocked Library
- 14:55New Relic
- 15:21Datadog
- 17:27Enterprise Runtime for Node Js
- 18:12Console Overview
- 19:10Load Times
- 20:34Conclusion
related talks
29mExamining Observability in Node.js
38mWelcome to Mordor! Hunting Performance Issues in Node.js.
01h 04mNode.js performance analytics and best practices
25mNode js applications diagnostics under the hood
57mTroubleshooting Node.js Applications Under the Hood
26mObserving Node.js: Using Metrics to Improve your Application Performance
39mWelcome to Mordor! Hunting Performance Issues in Node js
50mDebugging Node.js in Production