MPM and KeepAlive. Not a second copy of the NGINX cache page.
cPanel and some DirectAdmin stacks still serve through Apache. prefork with mod_php wastes RAM. Event plus PHP-FPM is the change, after a test account, not a global flip.
Workers, keepalive, and .htaccess that runs on every hit
prefork is required for mod_php. If the site can use PHP-FPM, event uses less RAM per connection. EasyApache is the path on cPanel. A hand edit is how the next upcp reverts it.
On for a site with many static files. Off or short if workers are stuck on idle browsers. The scoreboard says which.
A long rewrite file is parsed per request. Rules that belong in the vhost are moved there, on one account first. We do not delete a plugin’s rules blind.
Apache down is support. A slow PHP page may not be Apache.
If NGINX already sits in front, tuning Apache workers without looking at the proxy is the wrong layer. We check who owns port 80 before editing.
€55 an hour. An MPM change is agreed, not slipped in.
Minimum 1 hour. Switching prefork to event can break a site still on mod_php. That account is identified before the switch.
Apache using more RAM than the sites justify?
Tell us if it is cPanel and whether PHP is mod_php or FPM. No root password in the first message.