lighthouse4u-lambda

This is a sample project that can be cloned or forked to bring the power of Lighthouse4u and Google Lighthouse to AWS Lambda.

Architecture

Designed to be as simple as possible and embrace the serverless model, pay only for what you use, with near limitless scale.

AWS services required:

• Lambda - A single Lambda will be able to run Google Lighthouse jobs and supporting the Lighthouse4u UI+API. The consolidation is to keep things simple and reduce chance of cold starts, but can easily be broken into two Lambda's if desired.
• SQS - Simple Queue is where all Lighthouse jobs are stored until completed, and the Lambda will consume these events directly via triggers. More on this later.
• S3 - All reads, writes, and (limited) queries leverage S3 as the primary store.
• API Gateway - Attached to the Lambda to support the Lighthouse4u API+UI.

Prerequisites

• AWS Account
• AWS CLI installed locally, and configured to point to your AWS Account
• Docker installed locally
• Clone (or fork) this repo

Setup

Below are a set of one-time tasks to configure your environment. Be sure all services reside in the same region.

• Create a Standard Simple Queue with a default visibility timeout of 1 minute. Name it what you like, such as lh4u. You'll need the queue URL later.
• Create a S3 Bucket. Name it what you like, such as lh4u. No other steps are required on the bucket unless you wish to auto-delete older objects via Lifecycle Management.
• Create a Lambda Function. 1.5GB is recommended for optimal reliability. Name it what you like, such as lighthouse4u-lambda. Don't worry about deployment yet, just getting the shell in place.
• Attach your SQS queue to your Lambda via Trigger in your Lambda settings. This will forward any SQS messages to your Lambda.
• Create a REST API Gateway. Name it what you like, such as lighthouse4u-lambda-api.
• Create a Resource under your API Gateway, configured as a Proxy Resource, with resource path of {proxy+}.
• Under your new API Gateway Resource you'll create a method of type ANY and attach your Lambda via the Lambda Function Proxy option.
• Repeat the previous step by adding the same ANY method to the root of the API Gateway.
• Deploy your API Gateway, and create a new stage if your first time deploying this API. The name can be anything, such as the environment, ala production.

Note: Configuring permissions is outside the scope of these instructions. But effectively you'll need to permit your Lambda to read/write to S3 and SQS.

At this point everything is wired up and ready for deployment. Optionally you may opt to attach your own custom domain to avoid having to use the ugly API Gateway auto-generated URL, ala https://983dsdf.execute-api.us-east-1.amazonaws.com/production/.

Configuration

Edit your config/defaults.json file (or environment file if you wish to use NODE_ENV instead).

• Add your region name to store.options.region and queue.options.region configurations.
• Copy/paste your SQS URL to queue.options.queueUrl.
• Copy/paste your S3 Bucket name to store.options.bucket.
• Above are only the required settings. There are numerous options if you wish to take advantage/tune any other Lighthouse4u or Google Lighthouse options.

If you wish to deploy from your local enivornment (faster), edit your package.json file:

• Update the deploy:local-to-s3 script and add the desired S3 bucket in place of DEPLOYMENT_BUCKET. You can reuse your LH4U bucket if you prefer.
• Update the deploy:s3-to-lambda script and add the previous S3 bucket in place of DEPLOYMENT_BUCKET.
• Update the deploy:s3-to-lambda script and replace LAMBDA_FUNCTION with the name of your Lambda.

Deployment

Anytime you make a change to config or code, you'll need to:

• Via command line run npm run build. This will build your node_modules folder using a Lambda-compatible Docker image.
• Via command line run npm run deploy. This will deploy the zip file generated from the previous step to your S3 deployment bucket, and deploy that file to your Lambda. This step will only work if you completed the deployment configuration steps above. Otherwise you can manually upload the lighthouse4u-lambda.zip found in the root directly in your Lambda UI Console.