AWS Lambda, Chrome, Chromedriver selenium.common.exceptions.WebDriverException: "unable to discover open window in chrome"

Question

Any help would be greatly appreciated. Really feel like I am sooo close...and yet so far. Sigh.

I created a docker container using the following dockerfile.

I ran the docker container on my desktop and it worked fine.

I am using:

Selenium version: 4.2.0 headless chrome=102.0.5005.61

However, when I uploaded to ECR and then deployed to AWS Lambda as a container, I get the following error message:

class 'selenium.common.exceptions.WebDriverException' Message: unknown error: unable to discover open window in chrome

'''

FROM public.ecr.aws/lambda/python:3.9.2022.05.31.17-x86_64

ADD app.py .

ADD ChromeProfile  /tmp/ChromeProfile
ADD screenshots  /tmp/screenshots

RUN pip install selenium webdriver_manager chromedriver-binary-auto
RUN yum -y update
RUN yum clean all 
RUN rm -r /var/cache/yum
RUN yum -y install wget
RUN yum -y install gnupg
RUN yum -y install unzip
RUN yum whatprovides /usr/sbin/useradd
RUN yum install -y https://dl.google.com/linux/direct/google-chrome-stable_current_x86_64.rpm

RUN wget -N http://chromedriver.storage.googleapis.com/102.0.5005.61/chromedriver_linux64.zip -P ~/
RUN unzip ~/chromedriver_linux64.zip -d ~/
RUN rm ~/chromedriver_linux64.zip
RUN mv -f ~/chromedriver /usr/local/bin/chromedriver
RUN chown root:root /usr/local/bin/chromedriver
RUN chmod 0755 /usr/local/bin/chromedriver

RUN mkdir -p /etc/opt/chrome/policies/managed/
ADD Manifest.json /etc/opt/chrome/policies/managed/
RUN chmod -R 775 /etc/opt/chrome/policies/managed

CMD [ "app.handler" ]

'''

This is my python code:

'''

def handler(event, context): 
    try:      
        
        chromeOptions = webdriver.ChromeOptions()

        chromeOptions.add_argument('--headless')
        chromeOptions.add_argument('--no-sandbox')
        chromeOptions.add_argument('--start-maximized')
        chromeOptions.add_argument('--start-fullscreen')
        chromeOptions.add_argument('--single-process')
        chromeOptions.add_argument('--disable-dev-shm-usage')
        chromeOptions.add_argument('--user-data-dir=/tmp/ChromeProfile')
        chromeOptions.add_argument('--profile-directory=Profile 2')
        chromeOptions.binary_location = '/opt/google/chrome/chrome'
        
        driver = webdriver.Chrome(executable_path='/usr/local/bin/chromedriver', options=chromeOptions) <=== Error message shows up here.

        GetChromeInfo(driver)
        
        URLString = "http://www.google.com"

'''

I have tried playing around with a wide variety of chrome options. Can't seem to find the right magic combo. It always works on my desktop and fails in AWS Lambda.

I need to use the profile because of a popup that happens asking "Do you want to application instead?"

Answer 1

I tried it out with Alpine and built a python alpine image loaded with chromium driver and browser. I deployed the container to ECR and downloaded it to Lambda.

Life is awesome. It worked.

Something about AWS Linux 2 that didn't let chrome start. I tried a gazillion combos. If someone ever knows what I did wrong, please let me know out of curiosity. I am sure it is something so simple and stupid.