So just want to say I love Check_Mk for many reasons and has helped us a lot. So we are on version 1.6.08p2 On Ubuntu 18.04 We have notifications set up if our switches go from UP to Critical for ping? So I see the warning is at .2 and critical is at .5? Anyone know what that time frame would be and how long before it would trigger notification? Lost pinging for 30 seconds? 5 seconds? Minute? My guess on this is 5 seconds if I am reading right but wanted to ask. I can test by taking the uplink out and see how long before I get a text but just trying to figure out conversion or how long before it is triggered. Or how long ping has to be lost to send a notification? Hope this makes sense
Any help or tips would be great. Have our servers on a time delay if they go down and not pinging but donât on switches because would like to know if they go down and during down times just set notifications to off.
Thanks
In your picture you have the round trip time and if you are over the warning or critical threshold for defined amount of check attempts then you will get a notification.
For a host check i personally do most time 3 check attempts before i sent a notification. If your ping is a service check on my systems it need 5 check attempts before the notification is sent.
You can configure this with the rule âMaximum number of check attempts for host/serviceâ.
High latency means not down
But the check also goes critical if no package is received.
Thanks for info Andreas. So you say high latency means not down are you referring to anything in picture or my graph or just a FYI? I looked on switch and nothing went down but lost pings so check_mk sent notification. I just have the pings set at default? Warn at .2 and critical at .5 so when it lost pings for under a second if I am seeing this right? You are saying to up to 3 check attempts(learned something new with check_mk, I was using delay period time if something went down like server at 5 minutes for restarts) Since switch is a check on host will set for 3 times? Hope I am hearing you right. Thanks again. Makes sense and learned some new things. Brant
You can do a delay like you had done it with delay the notification or like i wrote with check attempts. Both ways are legit.
This are the two relevant rules.
This is how it looks like at a service.
The result is that it needs 5 attempts before i get a notification. The check interval is not relevant. This is a advantage over the delay setting. It checks 5 times irrelevant if you selected have a check interval of 1 minute or 5 minutes.
What makes this situation more complex is that you must calculate for yourself if the notification is sent quick enough if the check interval is set to a higher value or if you need to decrease the check attempts for some services to get a quicker notification.
