Cloudron Forum

Welp.. essentially loosing support for netcup vServers would be very impactful for us. We´d have to convince a lot of customers to move to rServers or another hosting service, not even mentioning our time spent on migrating their servers.
I guess this is one of those problems where nobody is to blame - I get netcup´s stance, but cooking up two cloudron versions every release circle doesn´t sound practical either.

I´ll send an email to netcup customer support, maybe they can be convinced to add AVX support if enough of us are complaining.

Hey All,

we are seeing recurring out-of-memory restarts on one of our WordPress (Managed) apps and, after some investigation, believe the app is not applying the package's own Apache worker limit. I would appreciate your help getting this corrected.

The container OOMs roughly twice a day. During the last event, apphealthmonitor logged the OOM at 2026-09-02 10:32:37 UTC, with CPU peaking around 540% and memory at 100%. Baseline memory is only about 0.5 GB, so the app is healthy until a short burst of concurrent requests hits it.

What we found:

Apache (mpm_prefork, mod_php) is running with the Ubuntu default MaxRequestWorkers 150, not the managed default of 6:

$ ls -l /etc/apache2/mods-enabled/mpm_prefork.conf
... /etc/apache2/mods-enabled/mpm_prefork.conf -> ../mods-available/mpm_prefork.conf
$ grep MaxRequestWorkers "$(readlink -f /etc/apache2/mods-enabled/mpm_prefork.conf)"
MaxRequestWorkers       150

At the same time, /app/data/apache/mpm_prefork.conf does contain the expected managed values (StartServers 2, MaxSpareServers 3, MaxRequestWorkers 6, MaxConnectionsPerChild 100), but this file is clearly not being applied to the running Apache.

With 150 workers allowed and each mod_php worker using roughly 200 MB with our plugin stack, about 20 concurrent requests are enough to exhaust the container and trigger the OOM. With the intended cap of 6 workers this could not happen, we think.

A plain restart and the memory change from 3 to 4 GB did not restore the 6-worker limit, so this looks to us like it needs to be corrected at the package or instance level (?).
The issue is reproducible, and we are more than happy to give you access or run any diagnostics you need.

Thanks very much for your help!

I can confirm that after a reboot my VPSs suddenly gained AVX support.
This is a relief, it spares me from loosing a lot of time migrating servers.

Wordpress sends out notifications of about anyting imaginable to the admin - which is fine, if you are managing only a couple of sites.
There are plugins to suppress some of these notifications.

Now, with managed Wordpress, core updates are handled by cloudron, so those notifications inside WP and the E-Mails it sends out are quite useless/borderline annoying.
It seems like you can edit the functions.php to disable those "nags".

Wouldn´t it be a good idea to implement this as a default for cloudron managed WP?

@maxkreg said in AVX Support in your VPS/Server:

I´ll send an email to netcup customer support, maybe they can be convinced to add AVX support if enough of us are complaining.

FYI: I didn’t get a reply from netcup for 3 weeks now. No reply as in no response aside from the usual ‘we have received your message’.
I think this is quite worrisome.

@girish Awesome, thanks for your efforts and quick response as always!