Expand (and Bump) Section `Tune PHP-FPM` in `Server tuning` manual page
I have just made a number of changes to the FPM worker configs for our nextcloud deployment, and we seem to have seen a significant increase in the speed of the applicaiton overall. We are using the official docker images via a compose setup. Over the last few years, I seem to recall seeing a great number of users stating that turning up the worker setting in their configs resulted in massive performance gains, and this seemed to be the defacto place to start suggested when a server was slow.
I think this section ought to be expanded, and given higher importance, on the server tuning page. Given how often it seems to come up as a resolution, and especially given that we are using the official images, whose default settings therefore must be quite low for our hardware, I think at least a small amount of information on relevant config items, their default values, and what impact they will have would be justified, and help many sysadmins improve their performance more quickly. Currently a link is provided for a guesstimate tool that can suggest some appropriate values for a given hardware configuration, but I would like to see more of an explanation, and for it to have higher importance - I'd certainly ay it should rank well ahead of adding a redis container to the pod, forcing HTTP2 on upstream proxies, etc., when considering the payoff versus difficulty implementing (few lines in an INI file is easier than deploying new containers, so should be considered first). It was also advantageous to know that sending USR2 to the parent process caused it to reload, AFAIK without dropping any existing worker connections, so this is particularly important given we can play with this setting without needing to restart, which is the more general case for other tuning options.
Thanks!