High CPU load on cpanellogd, various domains

[rhm.geerts]

Since a few days I get these kind of messages from CSF.

Account: accountname
Resource: Process Time
Exceeded: 23861197 > 1800 (seconds)
Executable: /usr/local/cpanel/3rdparty/perl/514/bin/perl
Command Line: cpanellogd - http logs for accountname
PID: 25912 (Parent PID:24887)
Killed: N

We get this for various domains.
I get the mails during daytime about 13.00 hours. The amound of seconds (23.861.197) is strange because that would mean this would already be busy almost a year?
So I guess that's something CSF is not reporting correctly.

I'm just wondering the following:
1.) Why is csf reporting so many seconds if this is an event which runs once or twice a day?
2.) Is the only way to stop this to use the tweak settings to change hours? Because I now excluded the hours between 08.00 and 23.00 top recent cpanellogd from running, but is bandwith log (which is set to every 2 hours) then still running, or will this be disabled too?
3.) Why did this not happen before? It just started about a week ago. Could this be due to a Cpanel update?
CENTOS 5.9 i686 standard – srv1 WHM 11.36.0 (build 14)

 

[rhm.geerts]

After changing the hours, I now get the messages starting between 00:05 and 00.10 which is reasonably ok.

However, it's always the same domains on which this happens. About 7 or 8 from the about 50 accounts present.
Could this point to a certain cause?
The logs don't show anything strange as far as I can see.

 

[rhm.geerts]

I see I'm not the only one, another thread was already posted with the same issue:
http://forums.cpanel.net/f5/cpanellogd-http-logs-exceeding-limits-325921.html

So this one can be closed, I will use the follow-ups in there.