Shipping a Node.js API to production isn't hard. Keeping it stable, secure, and observable at scale — that's where most teams stumble. Here are battle-tested practices from engineers who've been through it.
1. Structured Logging
Use a structured logger like Pino — not console.log — that outputs JSON. This makes logs queryable in tools like Datadog, CloudWatch, or the ELK stack. Always include a request ID on every log line for distributed tracing. Your future on-call self will thank you.
2. Graceful Shutdown
Handle SIGTERM and SIGINT signals. Drain in-flight requests, close database connections, unsubscribe from message queues, and exit cleanly. Kubernetes and serverless platforms depend on graceful shutdown for zero-downtime rolling deploys — ignore it and you'll lose requests on every deploy.
3. Input Validation at the Boundary
Validate every incoming request using Zod or Joi. Don't trust the client. Return clear 400 errors with field-level validation messages — your frontend team will thank you, and your database won't be poisoned by malformed data.
4. Rate Limiting and Security Headers
Add express-rate-limit to protect public endpoints from abuse. Use Helmet.js to set security headers (Content-Security-Policy, X-Frame-Options, etc.) in a single line. These are five-minute installs that block entire categories of attacks.
5. Health Check Endpoints
Expose /health (liveness) and /ready (readiness) endpoints. The readiness check should verify database connectivity and any critical external dependencies. These are required for Kubernetes, ECS, and every modern orchestrator.
Conclusion
Production-grade Node.js is less about code tricks and more about operational discipline. Logging, health checks, graceful shutdown, and input validation are the foundation. Build on them before adding any new features.
