Backend7 min read

Node.js

JavaScript runtime for server-side development - event-driven, non-blocking I/O makes it ideal for APIs, real-time apps, and microservices.

Node.jsJavaScriptEvent LoopStreamsRuntime

skill: Node.js category: Backend tags: [Node.js, JavaScript, Event Loop, Streams, Runtime] readingTime: 7

Node.js excels at I/O-bound tasks - APIs, streaming, real-time. For CPU-bound work (image processing, ML inference), use Worker Threads or delegate to a separate service in Python/Go. For type safety, use TypeScript from day one.

Workflow

Set up a Node.js project with TypeScriptTypeScript catches type errors at compile time and improves IDE experience - worth the setup cost for any project larger than a script.
Understand and use the event loop correctlyNode's single-threaded event loop means a blocked main thread blocks all requests. Never block with sync I/O or heavy computation.
Handle errors properlyUnhandled promise rejections crash the process in Node 15+. All async code needs try/catch or .catch() handlers.
Stream large data instead of bufferingLoading a 1GB CSV into memory crashes the process. Streams process data in chunks at constant memory.
Scale with cluster modeNode.js is single-threaded. Cluster mode spawns one worker per CPU core - fully utilizing a multi-core server.

Output Format

Production Node.js service with TypeScript, async error handling, streaming, and cluster mode.

src/
├── index.ts          # Entry point + cluster primary
├── server.ts         # HTTP server setup
├── routes/           # Route handlers
├── middleware/       # Error handling, auth, logging
├── services/         # Business logic
└── utils/
    ├── asyncHandler.ts
    └── gracefulShutdown.ts

Examples

Input

Output

Edge Cases & Gotchas

EventEmitter memory leak - Adding listeners in a loop without removing them leaks memory. Use emitter.setMaxListeners(n) or emitter.removeAllListeners() on cleanup.
Timer callback after process exit - Timers (setInterval) prevent clean process exit. Always clearInterval in SIGTERM handlers.
require() cache in long-running processes - Modules are cached after first require(). For hot-reloading config, delete require.cache[require.resolve('./config')] before re-requiring.
DNS lookup blocks - dns.lookup() uses the thread pool (not async DNS). Under heavy load, DNS lookups queue and cause latency. Use dns.resolve() which is truly async.

When to Break These Rules

Use Go or Rust for extremely CPU-bound workloads - Node.js Worker Threads help but aren't as efficient as natively compiled code.
Use Python for ML inference pipelines - the ecosystem (PyTorch, NumPy, scikit-learn) is incomparable in Node.js.

Reference Files

https://nodejs.org/en/docs/
https://nodejs.org/en/learn/asynchronous-work/event-loop-timers-and-nexttick