Is there any functionality in Nagios Core so that when a service or host alarms, and notification goes out, that if that server/host stays in a Critical or Warning state, that when the notification_interval comes up again for notification, that it can know it was already in that state previously?
Let me explain this a little better. Say a root filesystem service check on hostX goes critical at 01:00. The notification goes out, of whichever method I prefer. The notification_interval is set to say 1440 (1 day) and then at 01:00 the following day, another notification goes out. It goes out again because the state of that service/host hasn't gone OK (recovery). Can that second notification include some kind of parameter that I can script against that will let me know this is still in an unresolved Critical/Warning state?
If not, then I can simply script around it so that my notification script keeps a property file of sorts and then 1440 minutes later, if the property file still contains data, i know it's the 2nd or more occurrence. If a recovery occurs, I can remove that entry from my property file. This does seem like extra coding that if it is possible without and a parameter knows this; I'd love to KISS, so to speak.
Thanks.
Nagios notification_interval knowing previous state
Re: Nagios notification_interval knowing previous state
I think what you're looking for are escalations (which occur after x times, so this would work for your idea.) This would allow you to specify a different notification command to run, which could be 'notify-host-by-email2' for example, and contain the extra data. See this link - https://assets.nagios.com/downloads/nag ... tions.html
The other theory you mentioned on your own script would work as well, and your logic sounds spot on.
The other theory you mentioned on your own script would work as well, and your logic sounds spot on.
Former Nagios Employee
Re: Nagios notification_interval knowing previous state
Thanks, that looks like it could work. I guess I need to decide, since I have quite a lot of services defined, this will be just as many service escalations I need to create. Those examples give a lot of power to the escalations that I can see very beneficial for some people. I wish this could be implemented with less labor, perhaps wildcards for service names etc.rkennedy wrote:I think what you're looking for are escalations (which occur after x times, so this would work for your idea.) This would allow you to specify a different notification command to run, which could be 'notify-host-by-email2' for example, and contain the extra data. See this link - https://assets.nagios.com/downloads/nag ... tions.html
The other theory you mentioned on your own script would work as well, and your logic sounds spot on.
Re: Nagios notification_interval knowing previous state
You should still be able to assign a * for the members on an escalation.
The other option for large / dynamic scale, would be to run the route you mentioned with dynamic /tmp/ files, which would allow you quite a bit of flexibility / control.
The other option for large / dynamic scale, would be to run the route you mentioned with dynamic /tmp/ files, which would allow you quite a bit of flexibility / control.
Former Nagios Employee
Re: Nagios notification_interval knowing previous state
Just checking in since we have not heard from you in a while. Did @rkennedy's post clear things up or has the issue otherwise been resolved?
Former Nagios employee