We use a 3rd party tool to monitor some SQL Servers on our farm. After an upgrade we had a memory issue( leak ) on one of the nodes. Once the memory usage crosses 90% you can not re start the service seems the server needs a rebooting not an acceptable solution. If you just bounce the services prior to the 90% mark things stay fine and he product continues working as expected. We developed this as a temporary band aid until the bug fix is sent and applied. This is scheduled from Windows task scheduler. I hope someone may be able to make use of the alert.