Relevance to select two computers per location

Hi,
We have about 2300 physical locations with a varying number of computers at each of them. Each location has a unique property value like “LocA, LocB, LocC” so the list of computers looks like this:

LOCATION NAME	BES COMPUTER ID
LocA	                1001
LocA	                2002
LocA	                3003
LocA	                4004
LocA	                5005
	
LocB	                1002
LocB	                2003
LocB	                3004
LocB	                4005
LocB	                5006
	
LocC	                5001
LocC	                6002
LocC	                7001
LocC	                7002
LocC	                7003
LocC	                8001
LocC	                8002
LocC	                8003
LocC	                8004

We are interested in creating an automatic group that selects the two computers with the lowest computer id numbers at each of the 2300 locations to be used as “alpha” deployments for software updates (4600 computers total for the alpha ring). That way, as computers get replaced over time or might be powered off for some reason, we always have exactly two computers per location in the alpha group.

We know how to do this with an external script via the API, but we would really prefer to do it with relevance so we can keep everything inside of BigFix and not have to maintain the external script process. We also do not want 2300 individual groups created for the locations.

Does anyone have a suggestion for how to write this relevance?

Thank you,
Tom

I’m afraid it’s not possible with only client relevance - client relevance is evaluated at the client, and the client does not know the computer ID of any computer other than itself.

This would require something running on the server side, that has visibility into all of the systems’ computer IDs. So either a REST API script or a Console Dashboard. I wouldn’t spend the effort on new Console Dashboards at this point, given we are trying to deprecate the Console in favor of Web interfaces, so that leaves us with REST.

Thanks Jason. Appreciate the quick reply.
Tom