Cypress on AWS Lambda

This repo demonstrates how to run Cypress on AWS Lambda, executing each spec file in parallel.

The blog post, part 1 walks through the nitty gritty of getting a single Cypress spec running on Lambda (see branch part-1 for where this leaves off).

Part 2 contains details on invoking a Lambda function for each spec file and collecting the results in an mochawesome report.

Warning this is a prototype and it hasn't been battle-tested.

How do I use this?

Requirements: docker, terraform, nodejs >= 8

First build the image to gather dependencies

cd lambda
docker build . -t cypress-lambda

Extract dependencies from container

./deps.sh

Invoke Lambda handler locally

Verify that the handler invocation works when run locally with docker-lambda's: nodejs8.10 runtime, a dockerized replica of the AWS Lambda execution environment:

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

Deploy it

Back in the project root:

touch lambda.zip
terraform init
terraform apply

(You'll have to change a couple things in lambda.tf for your situation, such as the s3 bucket name, which is globally unique)

Invoke it

In the project root, first:

npm install

And then:

npm test

This will invoke the deployed Lambda function once for each spec file, collect individual mochawesome JSON reports for each Lambda invocation, merge together the JSON reports, and produce a single mochawesome HTML report.

Use your own tests

So right now this is using the cypress kitchensink example; you'll want to put your own tests in lambda/cypress/integration and then go through all the steps above again.