When aborting a pipeline, it tries for 10 seconds then starts going down the path of killing everything in sight.  It would be much better if there was a way to tell Jenkins that job abort takes longer than 10 seconds on a per job basis and be able to set a global upper limit on how long jobs may defer.