2.1.0p20 CPU load on server

[bjorns163 (Bjorn Stout)]

CMK version: 2.1.0p20
OS version: Description: Ubuntu 22.04.1 LTS

Error message:
I’ve been running 1.6 for a while now and decided to do a fresh install of 2.1 and migrate everything over to the new server. Now i got about 550 hosts and 6500 services but the CPU load on the server is through the roof CPU is at 100% almost all the time. it’s using 32 ghz cpu power. i cant recall the old server ever using this much load. i have a vmware cluster where this vm is running on for the monitoring. but it’s fully consuming 1 hypervisor of CPU load all the time…

if i check the top command i can see nagios uses between 30 to 45% cpu and the rest and python3 instanced that get spanned.

i changed the intensity of monitoring of some groups from 1min interval to 3 of 5 min intervals but load is still to much

CRIT: The current livestatus usage is 100.00% (!!), 20 of 20 connections used (!!), you have a connection overflow rate of 0.00/s (!!)

[jlagendijk (Jesse)]

We run 2.1.0P17 with 202 hosts & 10500 services. We do use distributed monitoring, our main CheckMK server that does about 190 hosts and ±9000 of the services has only 2 cores and 4GB ram, yet I see no high CPU usage. Not sure if it’s a bug in your version but it’s definately not normal. We also run on Ubuntu 22.0.4.1 LTS.

[bjorns163 (Bjorn Stout)]

trying to check if anything is of in the logs

web.log shows these entrys alot.:

2023-01-31 12:21:13,935 [40] [cmk.web 3374352] error executing configuration test ACTestMknotifydCommunicationEncrypted
Traceback (most recent call last):
  File "/omd/sites/DGL/lib/python3/cmk/gui/watolib/analyze_configuration.py", line 164, in run
    results = list(self.execute())
  File "/omd/sites/DGL/lib/python3/cmk/gui/plugins/wato/ac_tests.py", line 456, in execute
    config = self._get_effective_global_setting("notification_spooler_config")
  File "/omd/sites/DGL/lib/python3/cmk/gui/watolib/analyze_configuration.py", line 205, in _get_effective_global_setting
    return get_effective_global_setting(
  File "/omd/sites/DGL/lib/python3/cmk/gui/watolib/sites.py", line 815, in get_effective_global_setting
    return default_values[varname]
KeyError: 'notification_spooler_config'

apache/error_log seems full of this error message

[Tue Jan 31 13:23:35.311693 2023] [fcgid:warn] [pid 3971978] [client 127.0.0.1:42456] mod_fcgid: stderr: PHP Deprecated:  fsockopen(): Passing null to parameter #2 ($port) of type int is deprecated in /opt/omd/versions/2.1.0p20.cre/share/nagvis/htdocs/server/core/classes/GlobalBackendmklivestatus.php on line 196, referer: https://myhost/DGL/check_mk/index.py?start_url=%2FDGL%2Fnagvis%2Ffrontend%2Fnagvis-js%2Findex.php%3Fmod%3DMap%26act%3Dview%26show%3DDGL-WH%26zoom%3D165

[bjorns163 (Bjorn Stout)]

i ended up deploying 2 additional host now load is fine.