Silly “poem” aside, one thing that has always irked me is when people get in the habit of just retrying failed tests.
It may be that you get infrequent errors that result from timing issues within the code that need investigating but if we’re used to random tests failing and just hit retry, we’re catching but ignoring a potentially nasty defect!
Personally, if a test can’t be trusted then it should be removed until it can be fixed. Logging tickets can help this happen. If that means that you have lots of tests needing to be removed, then that is a clear indicator that you need to rethink the design and approach of your automated tests and pipelines. You’re clearly doing something wrong!