Storytelling with Performance Observers
summary
Shows how to use the Performance API and PerformanceObserver to collect marks and measures from a Node server using require('perf_hooks'), demonstrating garbage collection and custom marks arriving in a single observer callback. Reveals a hard buffer limit of 150 entries that must be cleared to avoid leaks, patches two timing modules to run in Node, and finds async_hooks produces lower overhead and better cross‑service tracing than raw PerformanceObserver data.
AI-written summary · about
chapters (10)
- 0:00Introduction
- 2:19Why are we here
- 4:12Tracing
- 6:34Performance Timeline
- 11:12Performance Buffer
- 11:50Buffer Clears
- 12:36Buffer Clear
- 13:15Marker
- 13:34Performance Observers
- 20:38Summary
related talks
16mCoding Session: Stunning Performance
22mExamining Observability in Node.js
01h 04mNode.js performance analytics and best practices
38mWelcome to Mordor! Hunting Performance Issues in Node.js.
28mMeasuring User centric performance
26mObserving Node.js
39mWelcome to Mordor! Hunting Performance Issues in Node js
29mExamining Observability in Node.js