I have this as an alert to monitor RAM utilization for servers:
mempools.mempool_free <= “10%”
This seems to work fairly well with the exception of the virtual servers with dynamic memory. I get the occasional notification that the VM has exceeded 90% of it’s RAM. However when I go look at the data it reports the server as having 4.5GB RAM instead of the max of 8GB and that it’s definitely using over 90% of 4.5GB RAM.
Is there a way around this? For now I’ve set the delay to 10 minutes and that helps but doesn’t completely alleviate the problem.