Prevent the 'failed' state when VMSS autoscaled to subnet namespace size
VMSS easily turn 'failed' status with following step and cannot recover without delete and re-create the VMSS.
Repro step is:
1. Create Vnet and subnets. The address space of subnet for VMSS is /27 (for example)
2. Create AppGW(Small/1instance, no publicIP) in Vnet
3. Create VMSS under AppGW
... set the Autoscale rule as Min/Max/default : 2/25/2
... In my case, I created Standard_DS1_v2 size VMSS
4. change Autoscale parameter to 16/25/16
5. change Autoscale parameter to 25/25/25
want to prevent the state. thanks,
note: this issue occurs when enable 'overprovision'. while it can prevent by disabling overprovision, it will be far much nicer if systematically prevent the issue by Azure side.