I think that maybe I found a solution, or at least a working round of this problem, at least it seems to work reliably for me.
Try setting the Max. instances under IIS Server -> FastCGI Settings to 1.
It seemed to me that only certain requests called the php-cgi.exe process, so that it became an outcast and started the processor, usually when updating a message. When reading other posts on this issue, one of them mentioned the Max Instances setting and set the default value to 0 or automatically. I wondered if this could have a good effect if everything is not as it should be. I suppose (but this is not really my area of expertise) if a particular request triggers a blocking process, so FastCGI just creates another, leaving the first in its place. Somehow it seems that only one instance allows PHP to move from locking, and the processor remains under control.
For servers with a high level of requests, setting FastCGI to just one instance may not be ideal, but, of course, it exceeds the delays that I received earlier. Used in combination with WP-SuperCache and WinCache, now it seems that it’s nice to compress now.
Gnasher
source share