Cluster

The `cluster` module forks worker processes that share a server port via OS scheduling (primary accepts or distributes connections depending on setup). Each worker has its own V8 heap and event loop — crash isolation is per process.

					import cluster from 'node:cluster';
import http from 'node:http';

if (cluster.isPrimary) {
  for (let i = 0; i < cpus().length; i++) cluster.fork();
  cluster.on('exit', (worker) => cluster.fork()); // respawn
} else {
  http.createServer(handler).listen(3000);
}
				

Sticky sessions are required when in-memory session state lives per worker. In containers/Kubernetes, horizontal pod scaling plus a load balancer often replaces hand-rolled cluster — same idea, better ops tooling.

On interviews: explain process vs thread model, when cluster helps on bare metal, and why orchestration duplicated its role.

Common pitfalls: in-memory caches per worker causing inconsistency; no graceful drain on worker restart; assuming cluster fixes CPU-bound JS (each worker still single-threaded for JS).

The trade-off is balancing simplicity, performance, safety, and operability — name which axis you optimized and what cost you accepted.

Checklist:

• One worker = one JS thread for user code.
• Plan session affinity or externalize state.
• Coordinate graceful shutdown across workers.
• Compare cluster to container replicas.