Run Google Chrome Lighthouse on AWS Lambda.

Since version 2.x, lighthouse-lambda has the same major version of lighthouse. For example, lighthouse-lambda 3.x will use lighthouse 3.x.

This README is for version 3.x. To see older versions, visit:

• 2.x: https://github.com/joytocode/lighthouse-lambda/tree/archived-v2

$ npm install lighthouse-lambda --save

// index.js

const createLighthouse = require('lighthouse-lambda')

exports.handler = function (event, context, callback) {
  Promise.resolve()
    .then(() => createLighthouse('https://example.com', { logLevel: 'info' }))
    .then(({ chrome, start }) => {
      return start()
        .then((results) => {
          // Do something with `results`
          return chrome.kill().then(() => callback(null))
        })
        .catch((error) => {
          // Handle errors when running Lighthouse
          return chrome.kill().then(() => callback(error))
        })
    })
    // Handle other errors
    .catch(callback)
}

You can use docker-lambda to test your Lambda function locally.

$ docker run --rm -v "$PWD":/var/task lambci/lambda:nodejs8.10 index.handler

You can use docker-lambda to install dependencies and pack your Lambda function.

$ docker run --rm -v "$PWD":/var/task lambci/lambda:build-nodejs8.10 bash -c "rm -rf node_modules && npm install"

$ docker run --rm -v "$PWD":/var/task lambci/lambda:build-nodejs8.10 bash -c "rm -f *.zip && zip lambda.zip -r node_modules index.js package.json"

• The file will be big (at least 75MB), so you need to upload it to S3 then deploy to Lambda from S3.
• You should allocate at least 512 MB memory and 15 seconds timeout to the function.

Same parameters as Using Lighthouse programmatically.

Returns a Promise of an Object with the following fields:

• chrome: an instance of chromeLauncher.launch(), remember to call chrome.kill() in the end.
• log: an instance of lighthouse-logger (only if you set options.logLevel).
• start(): a function to start the scan which returns a Promise of Lighthouse results.

lighthouse-lambda uses the Headless Chrome binary (stable version) from @serverless-chrome/lambda.

MIT