availability set for pip addressed VMs
Consider this scenario:
1. I have an application that requries more than 150 endpoints so I am forced to use a PIP address
2. As I'm using a PIP address, the options around load balancing (in any mode) machines in an Availability set is removed and I can only have single VMs
3. Those single VMs are subject to the possibility of planned and unplanned interruption.
Either, increase the number of endpoints per VIP allowed (and allow port ranges) or develop a form of load balancing and availability set for PIPs. Permanent, reserved PIPs should also be available.