pytest.mark.parametrize potential weird interaction

[michaelgmiller1]

We are getting error code 5 (no tests ran) on our test runners.

However, we have many more tests than runners, and no runners are running less than 10 tests. My suspicion is that we have a few slow tests with pytest.mark.parametrize - and these are not being split up across runners.

Does a test with pytest.mark.parametrize get split up into different groups?

[michaelgmiller1]

I tested locally, and it seems like the tests are split out onto separate runners. I suspect we are seeing the error code 5 simply due to edge overflow, where it's not possible to efficiently allocate tests to an extra runner.