Hello hgs and dbs!
I have been making several tests and noticed that the worst performance occured after a MySQL 5.7 to Maria DB 10.3 upgrade. This is the case with the superpromos.info site. This site is currently using PHP 7.3, but I also tried 7.0, 7.1 and 7.2 which made no difference.
The performance of another test site which I upgraded to 2.12.1 wasn't as bad as the superpromos.info site, but still pages were taking from 7 to 10 seconds to respond. I removed that test site but will install it again and upgrade it to 2.12.2 as suggested. Tried this one with PHP 7.0, 7.1 and 7.2.
I reinstalled the main site WITHOUT upgrade (2.8.3) and this one is pretty fast even with the "1000" external javascripts. It responds in less than 2 secs and will try to make it faster by loading the "1000" external javascripts from the footer as suggested and will remove the code.jquery.com. This one is using PHP 5.6. I would like to upgrade this one but since it's so big I need a "modification" on the /admin/pages/index.php file so that when I click on the PLUS sign to expand a parent category, it opens the subcategories in a new page like Ruud did a few years ago (I opened a different thread for this).
It's 1h30 AM here and I've been doing all these tests today so I'm going to bed now, but I will definitely install the 2.12.2 and send you the ftp details.
FYI, this is what my host responded about the slow superpromos.info after I asked them about the MariaDB upgrade:
When visiting the site and monitoring resource usage the "mysql" process would instantly start utilizing 100% of one your server's CPU available CPU threads.
Once the query, or request to the site has finished or closed this process would go down just as quickly. This behavior did not seem to have an effect on the other sites or
processes on the server however and leads us to believe that this sit in particular needs to be adjusted in the queries it is making to its database as other sites on the server
are not showing similar signs or issues. It appears that part of the query to database is to continually gather all tables and count them for a specific search. This is likely
the cause of your issue however if this is the cause this would be an issue that would need to be changed at the site level with how the site is coded to make database queries.