Issue Secondary IP performance

[Andrea Bampi]

I searched everywhere, not only here, but I wasn't able to find any reliable answer, so I kindly ask a help from the community.
I'm experiencing unexpected "performance" issues on a specific IP on a Plesk-managed server hosting about a dozen websites. Some of them have dedicated IPs, some others share the same IP (the main server IP, the same used for the Plesk panel).
The problematic site is a web application with an average of 10-15 users contemporary users (source: Analytics), secured with SSL.
I never had any issue before January, when I decided to get a SSL certificate and hence acquired a dedicated IP from my hoster; since January, I began experiencing frequent (meaning about twice a day for 4-5 consecutive minutes) "mod_fcgid: can't apply process slot for /var/www/cgi-bin/cgi_wrapper/cgi_wrapper" errors in the vhost error_log.
I know this isn't DIRECTLY related to Plesk, but (apparently) to the excessive amount of concurrent cgi processes (=users) ON THE SERVER, but the weird thing I noticed is that ONLY the aforementioned site experiences the issues (and becomes unreachable) - all other sites don't, they keep working without any issue (apparently not even slower).
So my question is - do you think it may have something to do with resource management/settings inside Plesk.. does Plesk assign limited/separate resources in my scenario?
Thanks in advance

 

[MarkM]

Don't think it has anything to do with the IP, more that the FcgidMaxProcesses limit has been reached.

This article will explain;

Some websites are unavailable: "mod_fcgid: can't apply process slot for /var/www/cgi-bin/cgi_wrapper/cgi_wrapper"

 

[Andrea Bampi]

Thanks for the answer, that page was the first one I found and checked when I first googled about the issue. The inexplicable thing (for me) is that FcgidMaxProcesses should fix an amount of concurrent processes for the entire server, but the only one firing errors and going down is the one I'm talking about. All other sites (12) keep working flawlessly. Am I missing something fundamental?

 

[Bitpalast]

I suggest to not only check the error_log, but access_log, too, at the time when the issue occurs. Sometimes malicious bots create a lot of useless traffic on a site that does not show up in statistics, but slows down the site plus causes overload situations like the one described.

 

[Émerson Felinto]

I suggest to not only check the error_log, but access_log, too, at the time when the issue occurs. Sometimes malicious bots create a lot of useless traffic on a site that does not show up in statistics, but slows down the site plus causes overload situations like the one described.

If this is the problem, setting the WAF on Plesk or enabling Cloudflare with the "medium-high" protection level should resolve.

 

[Andrea Bampi]

Thanks guys... I already checked the access_log file too, nothing alarming there (usual bot activity). My main question now is "Why does ONLY one site on the server (out of 12) fire the processing slot errors, while all other sites remain available (and without perceivable slowdowns, I'd add)?
I'm migrating the web application on another dedicated server now, but the main question still bugs me. Maybe I'm missing something about how the FcgidMaxProcesses variable works/is managed.

 

[Émerson Felinto]

It may be related to the number of open files your OS can load. Please paste here the output of the following command:

ulimit -a

 

[Andrea Bampi]

thanks for the replies, I've been busy on another project and I didn't have the chance to answer to the last suggestion. Anyway, I migrated the application on another newer dedicated server, so I won't investigate further.