CPU profiling

Learn CPU profiling and flame graphs.

CPU profiling

CPU profiling finds which functions consume your app's time. Node.js uses the V8 engine's built-in profiler. There are two paths: start with the --cpu-prof flag, or use the inspector module to profile at runtime.

The inspector module gives access to the Chrome DevTools Protocol. You open a session with new Session(), start the profiler, stop it, and read the result. This is a powerful tool for finding bottlenecks in production.

PerformanceObserver

PerformanceObserver is a class in the perf_hooks module. It catches performance entries (PerformanceEntry) live. You create it with new PerformanceObserver(cb) and start it with .observe().

For example you place performance.mark("start") and performance.mark("end"), then measure the gap with performance.measure("label", "start", "end"). The observer catches these measurements and passes them to the callback. This is a declarative, W3C-standard measurement method.

Exercise Log the type of PerformanceObserver

Import PerformanceObserver from perf_hooks. Log typeof PerformanceObserver.

Show example solution

const { PerformanceObserver } = require("perf_hooks");
console.log(typeof PerformanceObserver);

Test yourself

Which built-in module is used for CPU profiling?
- fs
- url
- path
- inspector
Show answer

inspector — The inspector module gives access to the Chrome DevTools Protocol.
Match each piece to its role

inspector

Profiling
Session

Connection
PerformanceObserver

Catches entries
profiler

CPU profile

Complete the module name

const { Session } = require("____")

Show answer

const { Session } = require("inspector")

Next lesson: Optimization

We Are Enhancing Your Experience

Cookie Preferences

Strictly Necessary Cookies

Analytical Cookies