How do you use the Log system to track test failures?

If it's not happening already, I would also suggest that someone familiar with your tests review the programming assignments.  We do this automatically as a part of scrum methodology, but it you are following a different process, try and insert QA early on, so you can see what's coming and then maybe know ahead of time that New Feature A is going to break Test 32 and allow time in the schedule to fix the test.