I see that adTempus has a mechanism to limit the number of jobs run concurrently in a queue.

Is there any way to set a limit on starting a new job if system utilization is above a specified amount? I would like to run as many jobs as possible, but not start another one until utilization falls below 90%. The utilization should be an average over the previous N seconds and should include CPU, memory, and I/O.

In terms of memory, I would like to run the system as hard as possible without paging.