Host Escalations

[rentsys]

The only problem I have with that fix is that if the host goes down, not unreachable, then the recovery won't be sent to robertsa.

[scottwilkerson]

then you may need something like this:

Code: Select all

define hostescalation {
        host_name                               TestUnknown124,TestUnknown125
        contacts                                rentsys,robertsa
        first_notification                      1
        last_notification                       0
        notification_interval                   0
        escalation_period                       24x7
        escalation_options                      u,d,r
        }

Escalation ranges CAN overlap
http://nagios.sourceforge.net/docs/nagi ... tions.html

[rentsys]

robertsa shouldn't be alerted when it is unreachable.
Correct me if i'm wrong but if a device has been escalated then it should only alert the contacts on the escalation when the device comes back up.

[scottwilkerson]

Now I may be confused to the desired behavior , but you can overlap escalation, whereas this would be valid

Code: Select all

define hostescalation {
        host_name                               TestUnknown124,TestUnknown125
        contacts                                rentsys
        first_notification                      1
        last_notification                       0
        notification_interval                   0
        escalation_period                       24x7
        escalation_options                      u,r
        }

define hostescalation {
        host_name                               TestUnknown124,TestUnknown125
        contacts                                robertsa
        first_notification                      1
        last_notification                       0
        notification_interval                   0
        escalation_period                       24x7
        escalation_options                      d,r
        }

[rentsys]

Alright, so imagine your the admin and you have a client.
You set up a PDU to be monitored for the clients.
Recently you have been having troubles with your internet provider and internet going in and out.
So you make the client not receive unreachable notifications.
The network drops and the PDU goes into unreachable because you set up parent child relationships correctly.
The network comes back up and the PDU sends notifications to the client that it is back up.
The client emails in asking if his device was down in the first place because he only go an up notification.
you tell him its fine, because it never actually went down.
That is the situation I am trying to avoid. The client only getting an ok notification.
if it's down first then the client should get the notification, but if it's unreachable first then the client shouldn't get a notification.

[lmiltchev]

This is an expected behavior. You get notification on recoveries, for both, down & unreachable states.

[rentsys]

I thought only the contacts that are escalated will get the alert.

[scottwilkerson]

I thought only the contacts that are escalated will get the alert.

Correct, this is why most people will add the original contact to the escalation as well as the new contacts. This is also described in the first paragraph of the Contact Groups section on page 2 of this document
http://assets.nagios.com/downloads/nagi ... ations.pdf

[rentsys]

But that isn't what is happening. The escalated contact 'rentsys' doesn't get the recovery alert. robertsa, who is not escalated, gets the recovery alert. Can you please set this situation up in a test environment? If it works then there is something wrong with my environment, and i can fix that. But it doesn't can you change the documentation because that is misleading.
All you have to do to set it up is make a parent child relationship between them, contact1 to both, then in a host escalation do this

Code: Select all

define hostescalation {
        host_name                               Host1,Host2
        contacts                                contact2
        first_notification                      1
        last_notification                       0
        notification_interval                   0
        escalation_period                       24x7
        escalation_options                      u,
        }

[lmiltchev]

robertsa, who is not escalated, gets the recovery alert.

The "robertsa" seems to be escalated...

Code: Select all

define hostescalation {
        host_name                               TestUnknown124,TestUnknown125
        contacts                                robertsa
        first_notification                      1
        last_notification                       0
        notification_interval                   0
        escalation_period                       24x7
        escalation_options                      d,r
        }