Each data source comes with its predefined set of thresholds, but most likely customer will desire to add new.
Now end user can do so if he/she knows the ITM attribute group and attribute name.
This is somehow putting back the ITM complexity in the scenario.
Customer will ask to create/modify a threshold straight from any graph in the detailed dashboards: "I see a graph for CPU usage and then with a right click I want to operate on threshold defined on that metrics". That way the attribute group and name will be pre-filled, and we don't ask customer to know them any more