0

I would like to know if it is possible configure Nagios 3.4.3 in this way: when a node is DOWN, its associated services will answer with not available (of course). However, in nagios web, selecting "Host Groups", I can see nodes with DOWN status but their services in CRITITAL status... It seems services have not been checked (of course because that systems are DOWN) but visually it seems services are in a problem because theey are CRITICAL.

This configuration, is a problem in my configuration or is it normal in Nagios?

Would it be possible to change status name (not CRITICAL) when node is DOWN?

thanks.

Improve this question

2 Answers 2

Reset to default
0

This configuration, is a problem in my configuration or is it normal in Nagios?

This is normal. It just means that the last service check that Nagios did, before it detected that the host was down, came back CRITICAL.

This CRITICAL service result may have triggered an on-demand host check, which resulted in Nagios discovering that the host is DOWN, at which point it begins to suppress notifications for the services.

Would it be possible to change status name (not CRITICAL) when node is DOWN?

Not really, no. Unless you consider the fact that it's open source, meaning anything is possible if you're willing to make some modifications.

It sounds like what you're really looking for is the various "filter" options for host/service state in the web interface.

Improve this answer
1
  • I can't say if before host was going DOWN, last check for that service was OK or CRITICAL. Is there any way for checking that? If I'm checking SSH, NIS client and NFS mount points services and my check command is check-host-alive (pings), is it necessary to create a dependency? Three services are important and it may be possible that SSH was down and NFS up (a problem with SSH), so I can't create a dependency of NFS and NIS to SSH. So I think that dependency should be of that three services for the ping (that is checking if host is alive or not)
    – Daniel Ruiz Molina
    May 18, 2017 at 8:26
0

It appears you have failed to configure your dependencies correctly. The services should depend on the host. There is also an inherits dependencies flag which may make dependency configuration easier.

It is common for one service to depend on another, often on a different host. These dependencies need to added.

Hosts may be unreachable due to failure of intervening routers/host/firewalls. These should be added to the configuration. Only the last hop dependencies need to be specified. However, you need to specify all dependencies. Nagios will respond appropriately to hosts/services that are unreachable/failing due to network/server/service outages..

It appears your service was CRITICAL before the HOST was discovered as down. This won't clear until the host recovers. You may be able to change its state by re-configuring it. The CRITICAL state is usually correct when the service's host is down. Use scheduled outage when the HOST is expected to be out.

Improve this answer
4
  • and what is the correct way for configuring these dependencies correctly? I didn't know it...
    – Daniel Ruiz Molina
    May 15, 2017 at 13:25
  • services have an implicit dependency on their associated hosts, by default
    – Keith
    May 15, 2017 at 15:52
  • I can't say if just before host is going DOWN, last check for that service was OK or CRITICAL... Is there any way for checking that?
    – Daniel Ruiz Molina
    May 18, 2017 at 8:19
  • @DanielRuizMolina It is possible to log all the requests at the server. This makes it easy to reconstruct the timeline in a case like this. I believe tests don't run if dependencies indicate they will fail.
    – BillThor
    May 18, 2017 at 22:42

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .