It is so comforting when senior senior senior engineers also tell you they have no idea why a fucking test spec occasionally fails. Literally, the spec is fine and all but sometimes, it just decides to say fuck off.

I would argue that spec is implemented badly, not knowing why it fails is the same as not knowing what it tests.

I say it is to broad and needs to be broken down, and/or the code is not proper testable.

Company I joined now had some of those, and I made a point of tracking the dam things down until they all passed consistently.

In all cases, they outlined bugs that could have occurred in production, but just hadn't yet through coincidence. It's worth investing the time to sort them out.

Ya'all have given me the motivation to sacrifice my weekend. I have been contemplating staying up through the weekend and figuring out wtf is going on. Well now ya'all have given me the juice. Will return with future updates.