Unless there is some huge misconfiguration of your base system one of the only reasons Galaxy could be causing this is that your users run demanding jobs and you execute them locally on the same node that Galaxy runs. Try using linux tools like ps or top/htop to identify what processes are eating your resources.