Question 1

What does the Node cluster module do?

Accepted Answer

Forks worker processes sharing server ports via the master, utilizing multi-core CPUs. Each worker has its own event loop/memory. Not a substitute for proper multi-host deployment, but useful on one machine.

Question 2

cluster vs PM2 vs Kubernetes replicas — how do you choose?

Accepted Answer

cluster/PM2 for process management on a VM; K8s for orchestration across nodes — prefer platform primitives in cloud-native setups. Don't double-cluster mindlessly (PM2 cluster inside K8s pods). One process per container is a common best practice.

Question 3

How does sticky session need interact with clustering?

Accepted Answer

In-memory sessions require stickiness or external session store; prefer external store. Load balancers and cluster round-robin complicate affinity. Design stateless workers.

Question 4

Worker died in cluster — how should the master react?

Accepted Answer

Log, alert, restart with backoff, and investigate crash loops (OOM, uncaught exceptions). Infinite rapid restarts can worsen outages. Capture exit codes and last logs.

Question 5

Lead: is Node cluster still relevant on Kubernetes?

Accepted Answer

Often less so — use multiple pods; keep one Node process per container unless profiling proves otherwise. Exceptions exist for specialized high-throughput hosts. Document the rationale. Prefer horizontal pod autoscaling.

Clustering Node.js Interview Questions

Other Node.js interview topics