CloudLinux gives a 503 error when EP limit is reached or when the user goes well over the top with CPU limit.
Litespeeds listens for 503 errors and does an automatic restart of the web server.
This means that when the server is using alot of ressources beacuse of a few accounts hitting their maximum allowed ressources Litespeed starts restarting the webserver.
Litespeed say to set CloudLinux's EP limit higher than Litespeeds Suexec max conn limit which is a pain as we currently have different EP limits per type of account.
I do not however see a solution for a user going well over his CPU limit…
Is there a way to change Cloudlinux's error message code to replace 503 Service unavailable with something like 408 Request Time out to prevent Litespeed from restarting their webserver ?
Also 408 sounds more user freindly that 503 that souds more link an internal error…