How do I deploy a Node.js API to serverless?

Deploying a Node.js API to serverless usually means packaging your API so it runs as functions behind an API gateway, instead of as a constantly running server.

Common options

• AWS Lambda + API Gateway
• Google Cloud Functions / Cloud Run
• Azure Functions
• Vercel / Netlify Functions
• Cloudflare Workers (best for edge-compatible code)

Typical approach

1. Make your API serverless-friendly

   • Avoid app.listen(...)
   • Export the app or handler instead
   • Keep startup code lightweight

2. Choose a framework or adapter

   • Express → use an adapter like serverless-http
   • Fastify → use @fastify/aws-lambda
   • Native handlers work too

3. Configure routing

   • Map HTTP routes to function endpoints via your platform config

4. Deploy

   • Use platform CLI / IaC:
     • AWS SAM
     • Serverless Framework
     • Terraform
     • Vercel CLI
     • gcloud, az, etc.

5. Test

   • Test locally with the provider’s emulator or local dev tool

Example: Express on AWS Lambda

1) Install

npm install express serverless-http

2) Create your app

// app.js
const express = require("express");
const app = express();

app.use(express.json());

app.get("/hello", (req, res) => {
  res.json({ message: "Hello from serverless!" });
});

module.exports = app;

3) Create Lambda handler

// handler.js
const serverless = require("serverless-http");
const app = require("./app");

module.exports.handler = serverless(app);

4) Deploy with Serverless Framework

# serverless.yml
service: my-api

provider:
  name: aws
  runtime: nodejs20.x

functions:
  api:
    handler: handler.handler
    events:
      - httpApi: "*"

5) Deploy

npx serverless deploy

Best practices

• Keep functions small and stateless
• Use a managed DB or external cache
• Reuse DB connections across invocations
• Minimize cold starts
• Bundle dependencies and remove unused code
• Put secrets in environment variables or secret managers

Important caveats

• Serverless isn’t ideal for:
  • long-running jobs
  • persistent websockets
  • heavy CPU workloads
• Cold starts can impact latency
• Local debugging differs from production behavior

If you want, I can give you a step-by-step deployment guide for a specific platform like AWS Lambda, Vercel, or Google Cloud.