We have jobs that fail because they're waiting for execution. In its current implementation, if you're even over capacity concerning build jobs (not that hard to be), the timeout step is completely useless.