Processes Basics

7 examples to get you started with Processes & Workers - 5 basic and 2 intermediate.

Prerequisites

• Node.js 24.18.0 on a multi-core host for cluster examples.
• Read How Node.js Works for single-thread JS model.

Basic Examples

1. process.pid and argv

Identify the running process and CLI arguments.

console.log('pid', process.pid);
console.log('argv', process.argv.slice(2));

• process.pid unique per OS process - log in structured logs.
• argv includes user args after node and script path.
• Container orchestrators identify pods by pid + hostname combo.

Related: os and process - host metadata

2. process.exit Codes

Signal success or failure to shells and CI.

function main(): void {
  const ok = true;
  process.exit(ok ? 0 : 1);
}

• Exit code 0 is success; non-zero fails CI and systemd units.
• Prefer natural exit after async work completes - process.exit skips I/O flush.
• Uncaught exceptions exit with non-zero automatically.

3. When to Stay In-Process

I/O-bound APIs should not spawn processes per request.

import { createServer } from 'node:http';
 
createServer(async (req, res) => {
  const data = await fetch('https://example.com').then((r) => r.text());
  res.end(data.slice(0, 100));
}).listen(3000);

• await yields during network I/O - no extra process needed.
• Spawning per request adds latency and FD pressure.
• Scale replicas horizontally before defaulting to workers.

Related: Event Loop Best Practices - keep main thread responsive

4. worker_threads Sketch

CPU work off the main JavaScript thread.

import { Worker } from 'node:worker_threads';
 
const worker = new Worker(`
  const { parentPort, workerData } = require('worker_threads');
  parentPort.postMessage(workerData.x * 2);
`, { eval: true, workerData: { x: 21 } });
 
worker.on('message', (msg) => console.log(msg)); // 42

• Workers share the same machine, separate V8 isolates.
• workerData clones serializable data at start.
• Prefer separate .ts worker file over eval: true in production.

Related: worker_threads - pools and SharedArrayBuffer

5. child_process spawn Sketch

Run external CLI tools as child processes.

import { spawn } from 'node:child_process';
 
const child = spawn('node', ['--version'], { stdio: 'inherit' });
child.on('exit', (code) => console.log('exit', code));

• spawn streams stdio - good for log tailing and ffmpeg.
• Avoid shell: true with user input - injection risk.
• See child_process for exec vs spawn.

Intermediate Examples

6. cluster for HTTP Fan-Out

Fork workers to use multiple CPU cores for accept loop.

import cluster from 'node:cluster';
import { availableParallelism } from 'node:os';
import { createServer } from 'node:http';
 
if (cluster.isPrimary) {
  const cpus = availableParallelism();
  for (let i = 0; i < cpus; i++) cluster.fork();
} else {
  createServer((_req, res) => res.end(`worker ${process.pid}`)).listen(3000);
}

• Primary manages worker lifecycle - restart on exit in production.
• Each worker has isolated memory - no shared in-process cache.
• Kubernetes often scales pods instead of cluster module.

Related: cluster Module - production patterns

7. Graceful Shutdown Hook

Stop accepting work before exit on SIGTERM.

process.on('SIGTERM', () => {
  console.log('draining...');
  setTimeout(() => process.exit(0), 2000);
});

• Coordinate with HTTP server.close() and DB pool end.
• See Process Signals & Shutdown for full pattern.

Stack versions: This page was written for Node.js 24.18.0 (Active LTS), npm 10+, TypeScript 5.6+, Express 5, Fastify 5, and NestJS 11.