It's a hard question but I can try to give a shot. I looked at our case database to see if some other clients had some issues and basically I saw, for some clients with around 10-15000 nodes on a Core Server, an increase of the DB size by 20GB essentially on table MP_ComputerPatchScan.
As you have already posted; here are 2 interesting articles - The MP_ComputerPatchScan table can grow quite large and Managed Planet Task Listener has not started/ What is this service used for
The best way to know the time it takes to fill it up, is to backup the LDMS DB and restore it in a test environment to give you a clear idea.
The performance hit will be when the service Task Listener will run, that's why it is advised to run it only when you need to generate your reports then stop after. You may schedule the report to run during the night indeed so.
I hope it will help you, but there is no much I could tell you than another client would report in a similar situation.