I was able to fix every domain on my list, and get it down to a swift 150MS on most of them, This was done my tweaking the following settings.

Major System Performance Items

– PHP-FPM (max children needed to be increased on busy sites) – NGINX RULEs  performance rules not enabled on many) – NGINX CACHE (caching issues with static not enabled)

Major Tasks I had to do

  1. Nginx Directive Bash Script (can’t be done in nginx.conf files, needed append in a vhost_conf.file in-between server tags ( this will be useful in the future to normalize the rules on sites. The rules I found differ from the performance rules we had – screenshot below (our rules on left) All the domains have been updated
  2. PHP-FPM script (plesk subscription settings were intermittent), pool calculations, averaging the amount of memory used per php instance, command of note below
    # gets avg memory per process of php-fpm in kilo bytes show in the RSS column 
    ps -ylC php-fpm7.4 --sort:rss 
    #php-fpm - max children calulcation = total servrer memory / avg memory per process = recommended children
  3. When using PHP-FPM , nginx caching is important, some of the domains that I found were not caching static images
  4. High DB Usage, This needs more investigation and conversation, the DB usage is too high. We have to look at multisite stack vs regular site stack (apache maybe better multisite etc) Lots there.

Domains to investigate 

Even though I was able to fix most of the sites we were working with(see below) tthere are still a number of ones causing problems.
  1. bodenseniorliving.com (little helps these one, it’s improved but still terrible avg 5000MS)
  2. eyesonplainville.com (same avg 8000MS)
  3. higginsvisioncare.com (same avg 5000MS) There are many more of these, can you find a commonality with these? They are acting like meadowcrest,compass, vistaeye but those were fixed from global settings I did. Once we know we can patch all the servers that same.
    Performance on after results