Host Downtime still trigger service notifications?

[jkearney]

I noticed that when I activate downtime for a host I'm still getting notifications for services on that host. Shouldn't the host downtime extend to all the services too?

I know I can choose to schedule downtime for all services on a host instead but I'd rather not do that because for some hosts that triggers a flood of 15-20 unnecessary emails which makes some of my bosses twitchy and irritated.

I guess I'm more just curious if this is a bug or if there's a problem with my install of XI.

[snapon_admin]

If the host doesn't go down, but the service checks all become critical then yes, you'll get alerts for the services. As you said, you can schedule downtime for all services and there is a checkbox when you do to "Schedule for host" as well. IMO, when a host is in downtime, Nagios should stop running checks on that host and the services associated with it, or there should at least be an option to do so when scheduling downtime. For me at least, if I'm scheduling downtime on a host it means that I'm aware that the host and/or the services associated with it will be down/critical during that time period. I see no reason that Nagios should even be continuing to use resources to check them if we're already aware that they'll be down. It also makes all of those checks show up as critical in some of the pages on Nagios (i.e. "All service problems" page) which can be annoying for the people that keep that page open or use that page to monitor for issues.

[jkearney]

If the host doesn't go down, but the service checks all become critical then yes, you'll get alerts for the services. As you said, you can schedule downtime for all services and there is a checkbox when you do to "Schedule for host" as well. IMO, when a host is in downtime, Nagios should stop running checks on that host and the services associated with it, or there should at least be an option to do so when scheduling downtime. For me at least, if I'm scheduling downtime on a host it means that I'm aware that the host and/or the services associated with it will be down/critical during that time period. I see no reason that Nagios should even be continuing to use resources to check them if we're already aware that they'll be down. It also makes all of those checks show up as critical in some of the pages on Nagios (i.e. "All service problems" page) which can be annoying for the people that keep that page open or use that page to monitor for issues.

I completely agree with everything you just said. So this is a not just related to my install which is nice to know. Is there a feature request area that we can post this to? Because this whole issue seems broken and/or very redundant and it would be awesome to have it fixed.

[snapon_admin]

I actually have an existing feature request for adding the ability to disable checks during scheduled downtime here: http://tracker.nagios.com/view.php?id=584. Otherwise yeah, tracker.nagios.com is where you want to go for feature requests.

[slansing]

You may want to post on snap-on's tracker thread, as well as create your own to add some weight. I've opened a internal feature request for this as well.

[jkearney]

You may want to post on snap-on's tracker thread, as well as create your own to add some weight. I've opened a internal feature request for this as well.

I just created my own and I'll also post on snap-on's. I feel like this should have been a no brainer to begin with but I guess not. Thank you!

[sreinhardt]

thanks for the feature request!

I feel like this should have been a no brainer to begin with but I guess not.

I would actually disagree. Nagios makes a very strong stance that you as an admin must tell it what to and not to do. Setting a host in downtime without services, should result in services going critical, notifications to be sent, and still should be checked. I have several use cases, where this is actually vital logic as simply having a host down, does not mean my services should stop.(they are largely oddball cases, but still) As for disabling checks entirely when in downtime. Again I have to disagree as the results feed directly into reports and graphs. Stopping a check entirely will result in perfdata issues, report inaccuracies, and potentially other problems. I do agree that there are cases that it could be useful, and is something we should consider adding, I just don't think this particular case was thought of.

[snapon_admin]

I actually agree with Spenser on this as well. I think there should be an option to disable checks, but I don't think it should be the only way it works. As he pointed out, I'm sure there are several cases where people would like to have checks continue to run or where the host being down doesn't mean that the services are critical. It just so happens that in my environment, and yours as well it seems, if the host is down then the services are broke as well.

[abrist]

To add to the reasoning for this decision, I just got out of a remote where this behavior was leaned on heavily:

They used event handlers to perform certain tasks when certain services were down - mostly starting services on other boxes. If the checks were not running, they would have not gone critical, thereby stopping the event handlers from performing their desired tasks. Just another use case for those looking for one.