Application Gateway needs load balancing methods
Currently it seems Application Gateway only allows a round robin of all servers in the pool - there's no way to change this.
Having a weighting of how many to each server, or picking the first of a list?
For an example scenario - Imagine two servers A and B and in general you want server A to handle all urls /projectA/... and server B /projectB/... for efficiency reasons. However both servers are able to serve up both sites if needed. You could have 2 pools, one with server A first and one with server B first on a "first-server" method. Then, if you needed to patch and reboot server A then during that time all requests would go to B untill A came back up. Then repeat for B
As far as I can tell right now you have to have 2 pools with 1 server in each. Then to reboot a server you'd need to add the other server to pool for a short time and remove it again afterwards.