Is it possible to reduce global polling frequency (across all sensors) from the default 1 minute to say 5 minutes?
This would bring down the number of false positives, flapping alerts etc. as well as strain on CMK server, clients and the network.
Data collected that way would still be useful enough to us.
Hallo,
which version of checkm (version and license) are you using?
And yes:
You can define rules but you should be sure that the reasons of the problems are ok.
Ralf
We are evaluating free edition 2.0.0p3 (CRE).
I’d like to experiment with different values and find the right balance.
By “rules” do you mean it has to be changed in many places, not a single one?
take a look here:
for example all DNS Check`s are running just every 1 hour
Greets Bernd
Hallo,
setup / host/ host monitoring rules.
a link to the manual The official Checkmk user guide
Thanks, it worked and appears to be a great improvement.
I have combined it with “Maximum number of check attempts for service” 1 → 3.
I don’t see dozens of stale services and hundreds of false positive emails:
OK → WARN, OK → CRIT
just to see
WARN → OK, CRIT → OK
a minute later.
It could be that CMK server would benefit from more CPU/Memory resources.
I couldn’t see any evidence of resource deprivation by running “htop” though.
take a look at your
or 
and also it`s a good idea to adjust the fetcher and helper like Andreas has writen here:
Thanks, that’s very useful.
Only one potential concern here: CRIT: The current livestatus usage is 99.04% (!!), 20 of 20 connections used (!!), You have a connection overflow rate of 0.00/s (!!)
Some discussion on that be found here: Analyze configuration Livestatus usage
This topic was automatically closed 365 days after the last reply. New replies are no longer allowed. Contact an admin if you think this should be re-opened.
