The IP addresses for your instances can be found by clicking on the Load Balancer resource and then “Backend Pools”. Azure will display the instances on the backend, along with their private IP addresses.
Simply RDP to the IP address of the instance and install the load simulator. Mess around with the settings and watch as the scale set adds and removes instances based on load.
If you ask me, the process of creating a scale set based on a custom image is a bit more tedious than it really needs to be. I suspect that, at some point, the functionality will be added to the portal. However, until then, we are left with PowerShell and ARM Templates as our options.
With that said, it never hurts to know your way around PowerShell and templates anyway so get to it. Go deploy your own autoscaling VM scale set using a template and add another skillset to your Azure arsenal!