We have continuously been getting an error when trying to run workflows.  The event log on the server shows that the error comes from having less than 5% memory available.  After we reboot the server, it works for several hours until the memory is under 5% available again. Has anyone else experienced this? If so, how did you fix it?  

We do not believe this is because of SQL since we limited that to use 12gb of memory (32 gb total in the server)