Hello, how do you implement something like "drain host after 10 consecutive failed jobs"? Unlike a host check script, that checks for known errors, I'd like to stop killing jobs just because one node is faulty. Gerhard