Configuration/optimization of the web server (OpenLightSpeed + ISP6)

Greetings esteemed gentlemen, excuse my rough language...

The point is this. There is a new VDS on Ubuntu 22 + OLS + ISP6pro - on it, 30 very different WordPress sites have been migrated from the old server (nginx+ISP4light). The new server is equipped with Object Caching Redis, and all sites are configured to work with OLS using the LightSpeedCache plugin. The server was prepared in a basic configuration by the Data Center itself.

It would seem that everything should work simply. But no... Cached requests (pages) are much faster, but with non-cached requests (pages, admin panel, Elementor) - it's a disaster and a mess. 😦 

For example:

🔸Old server: 6 cores / 12GB

cached (WPRocket) - 250...300ms, non-cached requests/pages - 3...5 seconds

🔹New server: 8 cores / 16GB

cached (LSCWP+Redis) - 25...30ms (!), non-cached requests/pages - 30...60....120 seconds (!!!) or may even result in a 503 error.

Simple one-page sites don't mind, they work quite quickly. But if the site is more or less loaded, not to mention WooCommerce - everything is just sad.

However, no matter how bad the site issues are, ISPmanager itself works fine without lags.

Load Average at rest is about 1.5...2, during peak up to 6...8, and both ISP and htop show load not exceeding 50% at peak. A long queue with an unloaded CPU suggests a MySQL issue. Also, MySQL is writing binary logs to /var/lib/mysql - over 200GB per day (!). Initially, during the site transfer, everything seemed normal, but towards the end, it became quite problematic.

There are a couple more questions about email operation, and some minor requests regarding bot protection, ...

I see no point in contacting the Data Center itself for configuration, as they start spouting nonsense about PageSpeed and CSS optimization...

Gentlemen, I urgently need qualified assistance.