Nagios Support Forum

The BPI groups determine state by looking down only one level. In the image below, should the status for the root group "Symphony" be set to Warning instead of Critical since both children 1-level down have Warning status?

Should the Group Health % for root node "Symphony" be 0% instead of 100% as shown in the image?

What the warning and critical thresholds are set at for BPI group "Symphony"?

Note: I merged your two topics as these issues are probably related.

The Warning and Critical thresholds are set for root group "Symphony" as follows:

Warning: 99%
Critical: 51%

On the first image where both Children of Symphony group have Warning status, the root group Symphony should also have Warning as the status since both Chilren are "Essential Members". With the children being Essential Members, the most severe status of the children should be reflected one level up.

On the second image where everything (root and children) are in Critical status, the health of the root Symphony should be 0%.

Would you please send us your BPI configuration? In my experience your behavior is unexpected, and I'd like to attempt a reproduction on our end.

Below is the config file without the localServices definition:

define sym-db-servers {
title=Symphony DB Servers
desc=
primary=0
info=
members=symdb1;Sym DB 1 - DB Conn Status;&, symdb2;Sym DB 2 - DB Conn Status;&, symdb3;Sym DB 3 - DB Conn Status;&,
warning_threshold=70
critical_threshold=40
priority=2
type=default
auth_users=
}



define sym-http-servers {
title=Symphony Web Servers
desc=
primary=0
info=
members=symhttp1;Sym Web 1 - HTTP Conn Status;&, symhttp2;Sym Web 2 - HTTP Conn Status;&, symhttp3;Sym Web 3 - HTTP Conn Status;&,
warning_threshold=70
critical_threshold=40
priority=2
type=default
auth_users=
}

define sym-backend-servers {
title=Symphony
desc=
primary=1
info=
members=$sym-db-servers;|, $sym-http-servers;|,
warning_threshold=99
critical_threshold=51
priority=1
type=default
auth_users=
}

It is possible that this is a bug (unless it is "by design"). I will have to discuss it with out developers. In our documentation, we have:

If an essential member has a problem, the entire group will be in a problem state, even though the thresholds have not been exceeded, and there is only one problem.

However, it seems like this "problem state" is always critical. I created similar BPI test groups and even if only one of my "essential" members goes into "Warning" state, the state of the parent group changes to "Critical" (not "Warning").

Please keep me posted. This is an important fix to release if it's not by design.

Our developers will be looking into this. There is an internal bug report filed already (TASK ID 5643) about this issue.

Does bug id 5643 covers both issues?

The group health % also seems off.

Thanks!