Docker

Docker is a powerful containerization technology that allows to encapsulate your test suite into a container that behaves the same on every system. This can avoid flakiness due to different browser or platform versions. In order to run your tests within a container, create a Dockerfile in your project directory, e.g.:

FROM ianwalter/puppeteer:latest
WORKDIR /app
ADD . /app

RUN npm install

CMD npx wdio

Make sure you don't include your node_modules in your Docker image and have these installed when building the image. For that add a .dockerignore file with the following content:

node_modules

info

We are using a Docker image here that comes with Google Chrome pre-installed. There various of images available with different browser setups. Check out the images maintained by the Selenium project on Docker Hub.

As we can only run Google Chrome in headless mode in our Docker container we have to modify our wdio.conf.js to ensure we do that:

wdio.conf.js

export.config = {
    // ...
    capabilities: [{
        maxInstances: 1,
        browserName: 'chrome',
        acceptInsecureCerts: true,
        'goog:chromeOptions': {
            args: [
                '--no-sandbox',
                '--disable-infobars',
                '--headless',
                '--disable-gpu',
                '--window-size=1440,735'
            ],
        }
    }],
    // ...
}

As mentioned in Automation Protocols you can run WebdriverIO using the WebDriver protocol or Chrome DevTools. If you use WebDriver make sure that the Chrome version installed on your image matches the Chromedriver version you have defined in your package.json.

To build the Docker container you can run:

$ docker build -t mytest -f Dockerfile .

Then to run the tests, execute:

docker run -it mytest

For more information on how to configure the Docker image, check out the Docker docs.