BPI Hosts

[BenGatewood]

Hi All,

I have created some dummy hosts to represent various BPI groups I have using the BPI Wizard. I was sure that if the BPI group went critical, the dummy *host* was supposed to go critical as well but that's not occurring. Only the dummy *service* assigned to the host is going critical. I tested this a few months ago and am pretty certain it used to set the host status as well. Have I missed something?

Thanks!

[WillemDH]

I've never saw the dummy bpi host go critical.

Grtz

Willem

[lmiltchev]

The status of the BPI dummy group would depend on the "health thresholds" you set up for the BPI group and whether the failing service is an "essential" member. Please, see our documentation on BPI here:

http://assets.nagios.com/downloads/nagi ... _Addon.pdf

[WillemDH]

Ludmill,

Yes, the bpi dummy group would indeed depend on the health threshold. But does this also count for the dummy host, created by the bpi wizard?

Willem

[lmiltchev]

When I was saying "bpi dummy group" I meant the one, created by the BPI wizard (not "dummy" bpi group, created under the "Nagios Business Process Intelligence" page -> "Create New BPI Group").

[WillemDH]

Hello,

Just made some availability reports and did notice the bpi hosts are on 100 % while, the 'root' service for the bpi group is 99.970 and 99.966 %. Any idea how to explain this if the host is supposed to take the status of the bpi group?

Willem

[lmiltchev]

The state of the service, attached to the BPI group (created by running the wizard) should match the state of the BPI group (created under the "Nagios Business Process Intelligence" page).

The BPI service needs to be attached to an object, so it's being attached to a "dummy" host. We care about the state of the service, not the host as the state of the service will match the state of the BPI group we are monitoring. I hope it makes sense.

For example, if I make (under BPI) a new group "test" and one of the services, which is set as an "essential" member, goes critical, the group will go to critical state as well.

When I run the BPI wizard, and select this particular BPI group, the service (BPI Process:test) will go critical, indicating that the BPI group "test" is in critical state.

The host itself will be in OK state as it's just dummy host.

This will not play very nice with the Availability reports as host's availability will be always 100%... but the state of the service is what we care about.

[WillemDH]

Hey Ludmill,

It makes sense, but I get questions from managers why the host is 100 % up in availibility reports, while in fact the bpi group, or in their beliefs the group is less then 100 %
Or is there a way to exclude the host data from availibility reports?

Grtz

Willem

[lmiltchev]

Or is there a way to exclude the host data from availability reports?

I don't believe this is possible. Do you absolutely need to use BPI groups in your environment? It's somewhat funny (in my opinion) that your managers brought up this issue. Usually managers are quite happy to see everything colored in green and 100% availability

[WillemDH]

Yea, but I need to represent a more representative view of the 'email' availability. (It's more a project leader then manager btw)

So just using host availibility won't cut it. As we use high availibility and F5 load balancers. Both nodes of a pair need to be down for the email availibility to be critical.. So I think I do need BPI groups for that to show no?

Grtz