Debugging Slow/Unreliable Skywalker Tests

Been writing some basic skywalker tests for a simple UI, mainly checking contents of a grid/map and then executing an action that happens on another test. I know 1. the test actually runs and 2. the UI passes the test, because it’s worked perfectly before.

Now the problem is that it works maybe ~20% of the time. All other times, it’ll just hang on one of the tests or fail everything (often, the error message will say the page/some component is undefined OR I’ll get a 500 Server Error OR something else). Naturally, this wastes a ton of time.

Generally speaking, am I supposed to expect such a low success rate for skywalker test execution (not passing the test, just getting it to run successfully)? If not, how can I go about debugging this?

tl;dr Are skywalker tests usually highly unreliable?

It’s possible that it’s slow because it’s doing lots of waiting. If one thing fails, then it might time out on subsequent tests.

In terms of unreliability, I’ve often found that something fails in my afterAll() and the test data isn’t cleaned up properly between tests. Try running the afterAll() in console, and see if it returns errors. Errors in beforeAll() and afterAll() generally aren’t reported back.