Does anyone know what controls the outbound ip address of a VM in an ARM deployment? The deployment I am working on has only a single public ip which is assigned to a load balancer. Some VM's are using this public ip as their outbound ip and some are not and there does not seem to be any logic in why. Ideally I want all vm's (which are a mixture of A nd D series) to use the same outbound ip, is this possible?
You can create an internet facing load balancer using ARM and Resource Groups. All the machines you want behind the load balancer will need to share the same backend IP pool.
You can find the steps to set one up here - https://azure.microsoft.com/en-us/documentation/articles/load-balancer-arm-powershell/