I'm running a big amount of tests (1000+), and all of them pass successfully, but eventually the build fails, without explaining why.

If I'm running each half of these tests separately, they both pass, and the build will succeed.
(which means the problem is probably not within the tests themselves, but rather the amount or something else)