HTML MASTER CLASS /// LEARN TAGS /// BUILD STRUCTURE /// SEMANTIC WEB /// HTML MASTER CLASS /// LEARN TAGS ///

Untitled Lesson

⚡ Total XP: 0|💻 backend XP: 0

Skill Matrix

UNLOCK NODES BY LEARNING NEW TAGS.

Select an unlocked node to view details root

Pascual Vila

Frontend Instructor // Code Syllabus

Common Pitfalls & Errors

The Error //

Capturing a CPU profile while the application is idle, with no requests being made

// Wrong: recording with no load generates nothing useful
$ node --inspect server.js
# (profiler recording, but nothing is calling the server)

// Correct: generate load WHILE recording
$ npx autocannon -c 10 -d 30 http://localhost:3000/endpoint

The Solution //

A profile only records data for code that actually executes during the recording window — an idle process produces an essentially empty, useless profile. Always generate representative load (ideally reproducing the specific slow scenario) against the application while the profile is actively recording.

The Error //

Assuming the function with the most "self time" in a profile is always the actual root cause

// A profile showing db.query() as expensive might mean:
// - It IS just slow (needs an index)
// - OR it's being called 500 times in a loop (N+1 pattern)
// The flame graph's call count/hierarchy reveals which

The Solution //

A function showing high self-time might simply be doing legitimately necessary work efficiently — the more useful signal is often which function is called an unexpectedly large NUMBER of times (an N+1 pattern) or which one has surprisingly high time relative to what it should logically cost. Read the flame graph's call hierarchy, not just the single highest bar.