This is a very complex subject, probably better suited to fora that discuss the internals of operating systems. Swap memory is not handled by applications, it is handled by the operating system (OS). Although applications can tell the OS which parts of their memory usage is more suited to swap storage, most do not.
You'll need to start by installing something that can monitor what is being swapped out to disk, at the very least, by application.
But before that, have a look at the memory allocations for all the services you are running on the server. Most services won't have an easy-to-see memory allocation, but Jira and Confluence do because they are run in Java. For both of them, have a look at their maximum heap size (the -Xmx setting in their environment. "ps -ef | grep -i xmx" will give you that for all the JVMs currently running in a Unix-like system).
If their values add up to more than the memory the system has available, then you can guarantee that bits of them are going to be swapped. And don't forget that you need to allow room for the OS as well - for "light" Unix-like systems, allow at least 512Kb (ideally 1Gb), and for most server or desktop distributions, 1-2Gb. If it's Windows, you probably need to allow 3-4Gb for the OS. So your sums start from "Jira + Confluence + OS space"
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.