I have a large set of tests running on a site. When I run the test locally on Windows - they all pass on 100%. The test are designed and running on Google Chrome.
Now, we have started to run the tests on Linux via Jenkins jobs on headless mode. Some tests now fail on 0% or only pass on 20% or even 10%. In my code I'm finding elements by ID, xpath or css and simple click on them. And I use the WebDriverWait object for waiting - both for the element to be present and to be clickable.
Example of my code:
WebDriverWait wait = new WebDriverWait(browser, secondsToWait);
wait.until(ExpectedConditions.presenceOfElementLocated(By.id(elementID)));
lastFoundElement = wait.until(ExpectedConditions.elementToBeClickable(By.id(elementID)));
clickLastFoundElement();
In my report I see mostly that the elements were not found and that I passed the timeout set in the wait object.
How to make headless tests be more stable?
Why the headless state causes so many problems?