The requirement is to facilitate the grouping of seemingly random machines - they’re not completely to the operator, but appear that way property-wise - without allowing it through the creation of manual groups, which are apparently the devil. For example, an operator may want to test the deployment of some new software to only the machines of the 15 local IT staff in his region. This may be a recurring requirement. A manual computer group makes sense for this, as it is easy to create; easy to target; and can be reused.
An idea would be to create a “Custom_Group” client setting, for example, that they could set on any target group of machines to make each machine part of that “custom” group. Problem is, that limits the operator to one group per machine (when normally, the machine has the option to be part of many Computer Groups). Similarly, one could permit the creation of a defined number of groups via client settings (e.g. Custom_Group1 - Custom_Group5, but it’s still not as easy manual computer group creation and would be more difficult to coordinate from the operators point of view (i.e. documenting which group settings are already used; and which setting/number corresponds to which group).
So that is the explanation of what it is that we’re looking for
All ideas are welcome!
@jgstew, @jgo - any additional thoughts?!