Flapping mechanism recognizes WARNING state on hosts?

Hi all,

we are currently experimenting with sending FlappingStart and FlappingEnd notifications.
I have configured notifications for hosts and some services.

The hosts are monitored via check_icmp, as are the services.

Over the night we got bombarded with emails, though there are no state changes visible in the history of the hosts:

My suspicion (hence the title):
Does the flapping mechanism recognize the WARNING state the check_icmp when the warning threshold for PL or RTA is surpassed, though there is no “real” state change on the host?

That would be fixed easily by just raising the warning threshold to the critical threshold.

Thanks and best regards :slight_smile:

Host states are mapped, 0/1 = UP, 2/3 = DOWN when returning from the plugin. Meaning to say, if the plugin exit status flaps between 1 and 2, you’ll have a lot of flapping start/stop involved here.

Coming from the other topic, I’m not sure why you’d want this type of notifications enabled for hosts … I would rather e.g. subscribe to the API event streams and “do something” with these events. On a second note, those notifications are also stored in the database, slowing down the table lookups.


Hi Michael,

thanks for the detailed reply!

Good to know! But it seems I have of an exit status flapping between 0 and 1, looking at the host history.

Sadly, it is not me who wants them but the service management responsible for the customer, as the customer also has a monitoring system in place, that sends these notifications. :confused:

Also good to know (for me at least). Will bring that point up and see if someone cares :smiley:

Will also try what effects changing the WARNING threshold has.

1 Like