Node.js theorytheory 0/50 · 0%
Performance · hard

45. Performance profiling

Finding the actual bottleneck instead of guessing.

Node includes a built-in CPU profiler: `node --prof app.js` records samples, and `node --prof-process` turns them into a readable report showing where time is actually spent. `perf_hooks` lets you measure specific sections programmatically.

js
import { performance, PerformanceObserver } from "node:perf_hooks";

const obs = new PerformanceObserver((list) => {
  for (const entry of list.getEntries()) console.log(entry.name, entry.duration.toFixed(2), "ms");
});
obs.observe({ entryTypes: ["measure"] });

performance.mark("start");
await heavyOperation();
performance.mark("end");
performance.measure("heavyOperation", "start", "end");

Profile before optimizing: intuition about "the slow part" is frequently wrong, and optimizing the wrong section wastes effort while the real bottleneck (often a database query or an N+1 network call pattern) remains untouched.

Check your understanding

  1. 1. What does `node --prof` produce?

  2. 2. What module provides programmatic timing marks and measures?

  3. 3. Why profile before optimizing?