Why do a bunch of CLOSE_WAIT processes hang on the server and it is unavailable?

Question

Nginx

Hello, my topic was earlier Why does the site give an endless connection wait?
The problem still remains: in the morning the site stops responding to clients and after a long wait 5** gives an error.
the site on php + nginx + apache + mysql and on cms bitrix, centos (bitrix VM)
this morning it did not respond again, and I decided to run the netstat command on the server and saw that at the moment I have 390 entries in the CLOSE_WAIT state.

where do they come from and it is because of them that the site is unavailable? (it’s worth restarting Apache and everything becomes ok and such a problem every day from 8 to 11 in the morning) (after 11 . I never fixed such processes at all and the output of the netstat command itself is 2 times less)
there are no errors in the logs at all, maybe it’s Bitrix with sessions that somehow starts to work wrong, what then needs to be fixed and looked in the configs for this, how to diagnose and defeat the problem?
I saw a topic with a similar problem https://searchengines.guru/showthread.php?t=780228

Answer 1

Apache has a service page with the path /server-status . You need to turn it on and see which workers hang on processing which link. And then dance from it.