AKS: default max pods && autoscale node group default subnet

[apetre070]

We ran into issues when updating the cluster to use Cilium network policy where both node pools with Max Pods = 110 does not leave enough room in the /24 subnet. This configuration situation would likely cause issues in other upgrade operations states as well.

The intent of these changes is to reduce that risk by:

1. Limiting the default max pods to 48 per node on both system and autoscaled node pools (This should still be adequate for default use cases).
2. Adjusting the autoscaled node pool to use a different subnet than the system node pool.

The result is that each node pool would use its own subnet and each subnet would have space for up to 5 nodes (10 nodes total by default).