Unfortunately, there is not a way to change this behavior. If I remember correctly(and it's been a while), we changed the behavior to make it easier for folks, since in most cases, users want to add more than just CPU and memory. Your workaround(select none, then only CPU) is currently your best option.
This change in behaviour managed to just bite us.
Operations staff weren't used to checking what was selected by default - adding a number of voice gateways with the (now) default behaviour added >500 interfaces per voice gateway.
It will be just as bad for switches.
I understand why some sites would like the new default, but I would like the option to go back to just CPU and Memory as the default...
How do I quickly detect devices that may have a large number of interfaces monitored?
I can educate, but it's going to happen again - I need some way to see when it occurs and on which devices.
So I can do a quick SQL Query that tells me which nodes have a large number of interfaces monitored.
SELECT Nodes.Caption, Interfaces.NodeID, Count(*)
LEFT OUTER JOIN Nodes ON Interfaces.NodeID = Nodes.NodeID
GROUP BY Interfaces.NodeID, Nodes.Caption
ORDER BY Count(*) desc
I really don't want to create another process/report however to keep an eye on this type of thing.
So i trained my users as i caught this after the upgrade. and it already hit me 2 times. It increases elemenets/load on the poller. It adds Vlan interfaces or loopbacks that we just dont need to monitor.
I don't like it either, Server 2008 R2 always shows 6+ interfaces and I know that it really only has one, so I have to click to change to none, then check off the ones I want. It's not that hard, but it a little bit of a pain now.
Can we request that this be logged as an enhancement request please (make the default behaviour selectable).
As a general rule, I would have thought that any change to previous default behaviour should be made selectable - to avoid situations just such as this.