Plesk & CloudLinux - Error 500 under load.

[jordan]

My RAM and CPU is only at 50% when connections are rejected under load.

[jordan]

Hey Guys,

I have a mystery problem with my server, after my server reaches a certain load approx 40/reqs a second it throws an error 500. My system reports that my resources aren even 50% used when connections are failing.

I am running static assets through an NGINX Proxy with the following configuration:

proxy_connect_timeout       600;

proxy_send_timeout          600;

proxy_read_timeout          600;

send_timeout                600;

gzip_proxied                any;

Applying the above config fixed error 500 has mainly fixed general problems like when a user interacts with a page. However, not under generalised load.

Does anyone have some ideas to diagnose and resolve this problem?

Kind regards,

Jordan

[bogdan.sh]

Hi,

Are we talking about static pages over NGINX only? But they are not limited by CloudLinux services...

[jordan]

Hi,

Thankfully when you mentioned "But they are not limited by CloudLinux services..." I had never thought about this, and checked out LVE.
Turns out that my entry process limit is capped at 20 and at the same point as the faults occurred my entry processes hit 20 as well.

I have now increased the limit, and the problem has been resolved.

Thank you very much for pointing this out,

Jordan

[bogdan.sh]

Then you was talking not about static pages

Good to hear you found the case here!