For some reason, there are a bunch of duplicate test cases (same identifier) which are being executed/generated.
I added some checks to the generator and there are ~90k duplicate tests.
Total test cases: 199853
Unique identifiers: 111385
This bug showed itself in the nightly reference test generator:

It seems that two instances of a test with the same identifier were being executed at the same time. The first one finished & left the second one in the queue forever, eventually timing out.
When debugging this, I noticed that this test showed up in the logs multiple times:
