I have updated my WP to 6.6 last week, along with a few other plugins that I got an update push. (I also run update&&upgrade for my Ubuntu 22.04 server on every weekends.) Now, the DB requires 10x longer time for every access. For example, on the admin’s posts list page, I needed 3~5 seconds w/ Redis Object Cache, ex-ante WP 6.6 update. Now I need 12~15 seconds, even if the ROC plugin is working behind. It’s about the same amount of time when ROC was turned off.
ROC seems working fine, based on ROC’s metrics and systemctl, and I can access DB from the server directly without any delay that I can see from WP admin (and real website).
I have checked with a few available tools if there is any other issues, but the DB, the ROC cache, and plugins are working all fine. as far as I can tell. I have several multisite networks on a single server (with remote DB), and all multisite networks show the same symptom. Since plugins are varying for each site, and some of them are based on different themes, the only explanation left to me is WP 6.6 update, but I do not have direct proof, thus I can’t find a way to fix.
Just in case, I also have set up a remote web server w/ all fresh installation of Nginx, PHP,… etc etc. The situation is the same. I don’t know what it is, but there is an unknown process that disallows ROC to perform it’s functionality.
I would like to know how to dig down and isolate the cause of this problem.
DB is MariaDB 11.2.4, Redis 6.0.16, ROC relying on Relay (v0.8.1) w/ PHP 8.2.21 and Nginx 1.27.0
I can also hear a lot more frequent fan noise from the server (I run an on-premise server)