Azure: Load balancer with pre-allocated IPs in backend pool

 When deploying a load balancer there are two different types of backend pool configurations: NIC based and IP based, see second screenshot below. When using IP based pools, you can pre-allocate IP addresses and then when VM's are added at a later stage they can be assigned the allocated IP addresses.

If you add VMs to the backend pool manually via the portal post deployment, the default choice is NIC based. If you want to deploy the NIC based setup with the ARM templates it's a bit more tricky. Then you have to deploy a couple of vNICs and then associate them with the pool. And then when the backend VMs are deployed, you can associate them with the already deployed NICs.

Note that there are some limitations around using IP based pools and private link services. According to this article "A load balancer with IP based Backend Pool can’t function as a Private Link service".

The content in this post uses this Microsoft article as a starting point.

The ARM template can be found on Github (lb-netw-dv-weu-001.json and lb-netw-dv-weu-001.parameters.json).

The LB is internal, has a frontend IP, one backend pool with two local IPs pre-allocated, one http probe and a load balancing rule that forwards requests on port 80, see more details in screenshots below.