I want to to tweak nagios in such a way that all checking stops (with services not displayed, or displayed as unknown) for any down node. Said differently I only want to see one alert for a down host instead of 1 (down) + n (1 for every service). Note that I am interested in service display/status, not only in turning off notifications.

Rationale: we use the nagios firefox/chrome plugin to monitor status and nagios' behavior is too noisy giving readings like these (because every node has 20 services):

3 down, 1 unreachable, 4 warnings, 87 critical

This means that the 7 critical services on up node (the problem is on the service) are swamped in a slab of red services which are critical only because they sit on a node that's down/unreachable. What I'd rather like to see is:

3 down, 1 unreachable, 80 unknown, 4 warnings, 7 critical

Or even

3 down, 1 unreachable, 4 warnings, 7 critical

I have looked in service dependencies but I did not fine a way to describe: "make all services on a alive-host dependen on the status of the host check".

I found the problem discussed here, where one of the participants thought it was a nagios bug, and here where one of the participants thought it was "as designed".

As things are, I am just interested in the effect, much less in the design philosophy. Note that this nagios is checking hundreds of nodes, so the maintainablilty of the solution is also important.

TIA and cheers.

  • It's by design. When the node does come up, nagios wants you to know if a service didn't come up with it. – Nathan C Jun 4 '14 at 12:07

You can hide this when using the latest version (0.9.11) of Nagstamon, which is a desktop (tray) status monitor for Nagios/Icinga: http://nagstamon.ifw-dresden.de/

enter image description here

| improve this answer | |
  • Cool. Turns out the (newest) Firefox extension has a similar filtering capabilities BUT Nagstamon also has additional one-click configurations/actions that make it preferable. Architecturally speaking I still think it should be in Nagios, but from a practical stand point this gets me 99% there. – Alien Life Form Jun 4 '14 at 16:16

You can do this by applying filters in the status CGI view...

The Chrome/FF plugins are just scraping data out of the CGI, so if you add these same filters to the status URL that the plugin is querying, you can control what it shows.

For example, if you only want it to see hosts: status.cgi?hostgroup=all&style=hostdetail

For problem services on hosts that are up, only: status.cgi?host=all&servicestatustypes=28&hoststatustypes=2

See this blog post on filters for more information about the specific values to use. (They're bit fields.)

Edit: this is precisely what nagstamon is doing when you check those boxes, btw.

| improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.