There is a new Increase & Optimize action in the panel. It scans the host and lifts the limits that usually bite under load — open files, the connection backlog, a few sysctl values — using one of three profiles.
It is reversible, and on a cluster it applies the same change to your nodes. If you have ever hit a wall at a few thousand connections and could not see why, this is very often the reason.