wukong-profiler

🔥 High-performance Node.js / CLI wall-time profiler, designed for real async / await workloads.

Key capabilities:

• ✅ Accurate async / await wall-time profiling

• ✅ Nested steps (true hierarchical Flame Tree)

• ✅ Automatic HOT / SLOW step detection

• ✅ CPU vs I/O heuristic classification

• ✅ Actionable performance explanations

• ✅ Chrome Trace export (Chrome / Perfetto)

• ✅ Profile diff for regression detection (CI-friendly)

English | 简体中文

📦 Installation

npm install wukong-profiler

yarn add wukong-profiler

Or use directly via npx:

npx wukong-profiler [options]

🧑‍💻 CLI Usage

npx wukong-profiler report ./profile.json

CLI

# Generate HTML report from a profile.json
npx wukong-profiler report ./profile.json

To profile code, use the programmatic API (createProfiler) or the examples in the examples/ folder.

🧠 Output Example

⏱ Total 28.52 s
├─ getGitLogsFast        957.78 ms  ⚠ SLOW [IO]
│   ↳ Likely I/O-bound (serial await or blocking I/O)
├─ getOvertimeStats      26.39 s    🔥 HOT  [CPU]
│   ↳ Likely CPU-bound (loops or heavy computation)
│   ↳ Deep call stack — consider flattening logic

Generate HTML report

# Generate HTML report from profile.json
npx wukong-profiler report ./profile.json

# Generate and open automatically in browser
npx wukong-profiler report ./profile.json --open

# Specify output HTML file
npx wukong-profiler report ./profile.json -o my-report.html

Usage: npx wukong-profiler report <profile>

<profile> is the path to a profile.json file (required).

Report subcommand options:

Note: This CLI only exposes the report subcommand. To profile code, use the JavaScript API (createProfiler) or the examples/ folder. For report-specific help run npx wukong-profiler report -h.

📘 Programmatic Usage

import { createProfiler } from 'wukong-profiler'

const profiler = createProfiler({
  enabled: true,
  flame: true,
  traceFile: 'trace.json',
  profileFile: 'artifacts/profile.json', // write profile to a custom directory
  hotThreshold: 0.8,
  failOnHot: true,
  diffBaseFile: 'baseline.json',
  diffThreshold: 0.2
})

// Example: write outputs to an `artifacts/` folder for CI
// const profiler = createProfiler({ enabled: true, profileFile: 'artifacts/profile.json', traceFile: 'artifacts/trace.json' })

profiler.step('load config', () => {
  // sync work
})

await profiler.stepAsync('fetch data', async () => {
  await fetchRemoteData()
})

profiler.end('Total')

Async / Await Profiling (Recommended)

wukong-profiler explicitly supports async profiling via stepAsync.

This guarantees correct wall-time measurement, even when the event loop is idle.

await profiler.stepAsync('getGitLogsFast', async () => {
  await readGitLogs()
})

await profiler.stepAsync('getOvertimeStats', async () => {
  await calculateStats()
})

Why stepAsync?

• ✔ Measures full async duration (not just sync part)

• ✔ Maintains correct nesting structure

• ✔ Enables accurate I/O vs CPU classification

🧩 API Reference

createProfiler(options)

Returns a profiler instance.

Options

| Name | Default | Description | | --------------- | ----------- | --------------------------- | --- | ------------- | -------------- | ------------------------------- | --- | ----------- | ------- | ------------------- | | enabled | false | Enable output & JSON export | | verbose | false | Verbose logging | | flame | false | Flame-style tree output | | slowThreshold | 500 | Slow step threshold (ms) | | hotThreshold | 0.8 | HOT step ratio | | traceFile | undefined | Chrome trace file | | profileFile | profile.json | Path to write profile JSON file | | failOnHot | false | Fail CI on HOT step | | diffBaseFile | undefined | Base profile for diff | | diffThreshold | 0.2 | Regression threshold |

profiler.step(name, fn)

Measure a synchronous step.

profiler.step('parse config', () => {
  parseConfig()
})

profiler.stepAsync(name, asyncFn)

Measure an async step with full wall-time accuracy.

await profiler.stepAsync('fetch users', async () => {
  await fetchUsers()
})

profiler.measure(name, fn)

Alias of step (sync). For async workloads, prefer stepAsync for clarity.

profiler.end(label?)

Finish profiling and output results.

profiler.end('Total')

profiler.summary(options?)

Get structured summary data for reporting or CI.

const summary = profiler.summary({ top: 3 })

summary.top.forEach((step) => {
  console.log(step.name, step.ratio)
})

📊 Profile Summary (Top HOT Paths)

const summary = profiler.summary({ top: 3 })

summary.top.forEach((step) => {
  console.log(step.path, step.ratio)
})

Examples

node examples/basic.mjs
node examples/flame.mjs
node examples/async.mjs

🧪 Chrome Trace

node examples/basic.mjs

Open:

chrome://tracing

Load the generated trace file.

or

https://ui.perfetto.dev

drag to Load the generated trace file.