Performance Profiler: Find the Bottleneck Before Your Users Do
**TL;DR:** Profile your Node.js application, identify bottlenecks, detect memory leaks, and measure event loop lag with flamegraph output.
The 10-Second Pitch
- **What it is:** A runtime profiler that attaches to your Node.js process and generates actionable performance reports.
- **Who it's for:** Backend engineers, SREs, performance-obsessed architects.
- **Why it matters:** Profiling in production is hard. This gives you production-grade insights without the overhead.
- **Stack:** Node.js (built-in inspector protocol), outputs JSON + SVG flamegraphs.
Setup Directions
Prerequisites
- Node.js >= 18
- A running Node.js service to profile
Step 1 - Install
```bash
npm install @mrtech/performance-profiler --save-dev
```
Step 2 - Profile
```bash
npx perf-profiler attach --pid=$(pgrep -f "node.*server.js") --duration=60s
```
Step 3 - Read Report
```bash
npx perf-profiler report --input=./reports/profile-123456.json --format=html
```
Pros / Cons
| Dimension | Rating | Notes |
|-----------|--------|-------|
| CPU Profiling | 5/5 | Native inspector, near-zero overhead |
| Memory Leak Detection | 4/5 | Heap snapshots with delta analysis |
| Event Loop Lag | 5/5 | Unique feature; directly measures JS main thread health |
Verdict
Performance Profiler is what you reach for when your p99 latency is degraded and your existing APM isn't giving you a clear root cause. The event loop lag measurement is the killer feature.
Skill: performance-profiler | Ecosystem: claude-code | Runtime: node | Auth: api-key