Analyse Configuration: Livestatus usage reports critical: where to increase livestatus connections

CMK version: 2.0.0p17 (CRE)
OS version: Ubuntu 20.04 LTS

Error message: CRIT: The current livestatus usage is 99.41% (!!), 20 of 20 connections used (!!), You have a connection overflow rate of 0.00/s (!!)

We are running a standalone CRE instance (just master, no multiside) and the “configuration analysis” tells us, that our livestatus is at its limit.
The information text does say:
"Livestatus is used by several components, for example the GUI, to gather information about the monitored objects from the monitoring core. It is very important for the overall performance of the monitoring system that livestatus is a reliable and performant.

There should always be enough free livestatus slots to serve new incoming queries.

You should never reach a livestatus usage of 100% for a longer time. Consider increasing number of parallel livestatus connections or track down the clients to check whether or not you can reduce the usage somehow."

But it does not give a hint, where we can increase the number of parallel livestatus connections.
The online docs does not inform, where these configuraitions are done.
Searching “livestatus” in the global settings seems also not helping:
2022-01-12 085812

Does anybody know, what to do to increase the number of connections for livestatus?

Hi Christoph,

If you do not use distributed monitoring and you have only one instance you dont need livestatus at all.
Check in WATO in distributed monitoring that the connection is setup with “connect to the local site”

on CLI you could disable livestatus with "omd config " → Distributed monitoring → Livestatus_tcp → OFF
You need to stop the site to do this changes

20 connections on livestatus for a single site looks suspicious. May you check with netstat
netstat -a | grep 6557 who s consuming this connections.



Hi Mike,
thank you for your help.
I checked the omd config and livestatus_tcp is OFF.

I called the netstat command and found nothing, meanwhile the critical alarm is gone (as I restarted the analyse again).

This is somehow weird - as this alarm seems to come periodically and gone by itself.
During the alarm I checked the config with omd config show and there livestatus_TCP was not OFF, but “LOCAL”.
Nobody has changed these settings - do you know the status “Livestatus_TCP = LOCAL” and which process set this during runtime?

Hello Christoph,

I checked on two standalone sites, one 1.6 and one 2.0. On both Livestatus is OFF.
I dont know the option LOCAL. In case Livestatus is not used over TCP it uses a local UNIX socket.

Where does the “Error message” comes from? Is this from a notification or did you found it in a log?



it’s inside the analysis page:

Both config and netstat does not show anything.

This message is perfectly normal on the RAW edition. There you get every time the “CRIT” message as this checked part is not existing inside the RAW edition. A little bit text here would be fine that this check is not relevant for RAW edition users.

1 Like

What a bunch of comedians :wink:

1 Like

thank you all for your help and the clarification.

In a perfect world, this check would be just invisable or disabled in the RAW edition, which would have saved some lifetimes.
Or a short remark in the explanation text.

Still funny, that the system does check something sometimes, as this check is sometimes “green- Ok” and sometimes “red - critical”.

I have disabled the check now, which works perfectly :wink:

1 Like