also, its worth noting that we found process monitoring on high activity systems to require a

--watchdog_level=-1

+

--watchdog_memory_limit=300