AWS Lambda

Run Curveball on AWS Lambda

Curveball runs natively on AWS lambda, but we do recommend getting some experience on localhost first with Node or Bun.

Before you start with this, it’s also important that you know on a high level of AWS Lambdas work on a high level, and how they respond to HTTP requests using the Node engine.

AWS lambdas use Node, but they don’t use the http stack to handle HTTP requests. These lambda functions export a ‘handler’ function that receive an AWS specific object with the HTTP request information, and are expected to return another object with the respone information.

The AWS documenation has a good tutorial that demonstrates a “Hello world” with just the Lambda API.

Once you understand the basics, this is how Curveball integrates with Lambda.


Get the NPM packages

npm install @curveball/kernel @curveball/aws-lambda
npm install --dev @types/aws-lambda

Hello world

This is the canonical ‘Hello world’ application in Curveball on AWS Lambda. Usually this would be called index.ts, but you are responsible for tunring this into index.js before uploading this to AWS.

import { Application, Context } from '@curveball/kernel';
import handler from '@curveball/aws-lambda';

const app = new Application();

// Add all your middleware here!
app.use( ctx => {
  ctx.response.body = {msg: 'hello world!'};

exports.handler = handler(app);

In Curveball (like Koa) everything is a middleware, so to respond to a request you use the .use() method.

When this method gets called, you are given a ctx argument. This is an object that contains all the information about the HTTP request (ctx.request) and the response (ctx.response).

In this case we are setting the reponse body to contain an object {msg: 'Hello world!'}. By default this will be turned into a JSON string.

The last step is that we export a function called ‘handler’, which AWS will be call for every request.

Now we’ve covered the Lambda-specific documentation, go read next steps to go over all the common use-cases.