We have a handful of servers that constantly run at a high memory used % throughout the day and then fall below warning thresholds in the evening. Everything is operating as expected and doesn't need to be fixed. The nuisance is that we want to monitor, track, and trend memory usage but we don't need to be notified if it goes high and we don't want to have to acknowledge the servers that spike daily.
For the first problem I just turned off notifications for the memory service across the board (it's applied to the hosts via a hostgroup anyway), so that was easy.
For the second problem, how can we either auto-acknowledge or not require acknowledgment for certain services? We don't want to see these as "unhandled issues" when it's a known, OK condition for the server to be in. I guess I could use the NRPE AlwaysOK script but I still want to be aware of high memory in the environment, but just have it accepted as a known condition.
How to not require acknowledgments for certain services
Re: How to not require acknowledgments for certain services
I feel like you have sort of answered your own question... the only reason for acknowledgement is to suppress alerts which you achieved by turning off notifications for those services. Though I suppose there is one exception: if you are doing ITSM reporting or something similar.
One way I can see of getting around this is if you assign those servers a new memory usage check where the warning and critical thresholds are over 100%, the check_nt plugin doesn't validate the range of the input.
The only other way of handling this in a more general sense would be to have an event handler for those services that will take the state information as input and then pump that back into the Nagios command file with a state of OK.
I don't think there is a friendlier Nagios way of setting an exemption.
One way I can see of getting around this is if you assign those servers a new memory usage check where the warning and critical thresholds are over 100%, the check_nt plugin doesn't validate the range of the input.
The only other way of handling this in a more general sense would be to have an event handler for those services that will take the state information as input and then pump that back into the Nagios command file with a state of OK.
I don't think there is a friendlier Nagios way of setting an exemption.
Re: How to not require acknowledgments for certain services
Well as I mentioned above NRPE has an "AlwaysOK" script will then fire off any other check you want, but they will always be "OK" regardless of warning or critical. While this would give me a pretty, all-green view, it's not exactly what I'm looking for. I want to see the yellows and reds, but I don't want it showing as "unhandled" problems.
I know I'm asking for something out of left field and it may not be possible. I just wanted to ask before I assumed.
I know I'm asking for something out of left field and it may not be possible. I just wanted to ask before I assumed.
Re: How to not require acknowledgments for certain services
Oh I see, the only way I could see being able to do that specific functionality is to use an event handler that will automatically acknowledge the service. Outside of that I don't see any other option to do exactly that.
Re: How to not require acknowledgments for certain services
I have to be honest that I don't fully understand event handlers. Is there an article or document you could point me to that could assist with an auto-acknowledge handler?