Hello everybody,
I have noticed that when rolling out changes in the Icinga Director always about 30% of the checks jump to unknown and after about 1 minute again get the status OK. Has anyone here a hint how I can eliminate the error. Thank you for your help
Used System
RHEL 7.7
Icinga2.11.0
Director 1.7
Icingaweb 2.7.1
Hi,
can you please add more details about the checks which switch the state to UNKNOWN? What says the output? Are these checks ran on an agent by any chance?
For agent checks it is recommended to have a dependency:
https://icinga.com/docs/icinga2/latest/doc/03-monitoring-basics/#dependencies-for-agent-checks
Best regards
Michael
Hello,
After the change to a service or service template via the Icinga Director, this phenomenon arises.
For information: The system consists of a master - master configuration as a cluster
for example:
I change the name of a service (add spaces at the end) and click save
Then I click on Activity Log and then roll out existing changes.
After about 10 seconds about 1500 checks jump on Unkwon (These are checks that are performed by the Icinga Agent).
After about another 20 seconds, the checks go back to the “OK” status step by step.
The Agent Check is configured on the Director Master of the following dimensions:
apply Service “Icinga Endpoint Connection” {
//display_name = “Icinga Zone”
check_command = “cluster-zone”
check_interval = 30s
retry_interval = 5s
max_check_attempts = 240
vars.admin_only = true
assign where get_object("Endpoint", host.name)
ignore where host.name == NodeName
}
Maybe a tip here for me …