Nagios UP, Access Switch down, no Notification should send

[zaji_nms]

Dear Expert

Yesterday our one of Access Switch was down due to loose power cord where our Nagios was connected to it. As Nagios server was up and running and when its unable to poll the devices (unable to ping customer router) generated many Notifications to our customer that Hosts are down.

We were in very bad situation to tell the customer please ignore its false alert.

Me sure there is some sort of Notification security to avoid such situation in the future that if Nagios unable to ping our last network (exit) point (in the diagram router 'Z') do not send any notification (just show the alerts) in the Operations Center.

We will configure some where in the Nagios the last exit point (Router Z IP-Address).

Regards

[jdalrymple]

This is very easy and well documented. It's called reachability logic and is all configured in the parents directive:

https://assets.nagios.com/downloads/nag ... ility.html

If you prefer service dependencies are also a valid option, where that your customer's services (probably better defined as a servicegroup for simplicity's sake) are all dependent on a check_ping to Z.

https://assets.nagios.com/downloads/nag ... ncies.html

I would recommend the reachability logic. If you had it in place Nagios would have drawn the map for you.

[zaji_nms]

Dear jdalrymple

I have gone through however there is many options. Will you please spare some minutes to guide bit more.

We have below Hosts which are scattered all over the world but within our network well known exit point is 10.10.2.39 (common in all traceroute result)
Host A
Host B
Host C
Host D
Host E
Host F

We have created Host nagios-exit-point = 10.10.2.39 (which we saw in all our trace result). Before sending any notification, Nagios should check 10.10.2.39 is reachable or not, if reachable then only do the next action (send notification) otherwise not.

so how to configure below options on Host Dependency Management via CCM

Inherit Parents
Execution failure criteria
Notification failure criteria
Manage Hosts
Manage Dependent Hosts

Regards

[lmiltchev]

You may have something like this (just an example):

Code: Select all

define hostdependency {
       dependent_host_name           		Host A,Host B,Host C,Host D,Host E,Host F
       host_name                     		10.10.2.39
       inherits_parent               		1
       execution_failure_criteria    		u,d,
       notification_failure_criteria 		u,d,
       dependency_period             		24x7
}

where, the "host_name" is the host that is being depended upon (the "master" host), and the "dependent_host_name" hosts are the "dependent" hosts. The "u" and "d" options under the "execution_failure_criteria" and the "notification_failure_criteria" directives indicate that the dependent hosts will not be actively checked (and notifications won't be sent) if the master host is in either an UNREACHABLE or DOWN state.

For more info, read the documentation ("Help & Documentation" drop-down menu) under the "Host Dependency Management".

[zaji_nms]

Thanks lmiltchev

Looks there is small bug , in Execution failure criteria even we selected and saved option = d (showing in View Text Config too) but when we revisit and showing config in CCM its got unchecked (if we blindly save again, option 'd' will get unchecked).

Nagios XI 2014R2.0

Regards

[tmcdonald]

I am not able to replicate this on XI 5, so I believe this has been fixed in later versions. I would suggest updating to at least 2014R2.7 if possible.