One of our agents has a job that runs every 5 seconds, which seems to stop for no apparent reason. There are no exceptions being thrown, another scheduled task, a heartbeat that confirms that the agent is running correctly, still works, and there are no error messages in the windows event log. The scheduled task runs correctly for a few days, but then stops for no apparent reason, as if the scheduler forgets to start the task again.
I have no debug information, since there are none, so my question is if anyone else has experienced this? Can it be caused by a scheduled task running for too long, or taking up too many resources? As in, the scheduled tasks don't terminate correctly (for some reason) and then the scheduler being unable to allocate a thread to start a new task?
There is a parallel discussion about this on the NServiceBus Google groups: https://groups.google.com/forum/#!topic/particularsoftware/0GAI9oqmp74