No, this is the aggregation function. This will determine what state the aggregation takes.
There are three types of aggregations:
-
Best - the node takes the best state of all its branches. Example: You have three load balancers. The aggregation looks across those three and picks the best state of the three. So if 1 is OK and 2 are critical, the node will be shown as OK
-
Worst - the node takes the worst state of all its branches. Example: You have three load balancers. The aggregation looks across those three and picks the worst state of the three. So if 1 is CRIT and 2 are OK, the node will be shown as CRIT
-
X out of Y - you can determine what share (in percent or in absolute terms) you deem ok. For the load balancer example, you may want to say: If 2 out of 3 are OK, the node should be OK, if 1 out of 3 is OK, you want it to go to WARN, if 0 out of 3 are OK, the node should be shown as CRIT
Now you can translate this aggregation logic to your host group. What is the “summary state” of this host group? Is it best (as long as one host in the group is ok, the group is ok) or something else.
Note: This will of course not give you the arithmetic average of host group. For example, if you have one host with 100%, one with 95% and one with 75% availability → the BI aggregation will not be shown as 90% available (which would be the arithmetic average). It will show you 100% if you chose “best”, 75% if you chose “worst”, and 95% if you chose 2 out of 3.
Does that make sense?