There is no documentation out for HA in WVD. The service itself has HA in-built, so you don't have to do anything if there are failures on the service side. But if there are issues in the region where the VMs are built, you can use Azure Site Recovery to build fail-over VMs in a backup region. Windows Virtual Desktop recommends using Azure Site Recovery to manage replication for Azure VMs replicating between Azure regions
How can I use ASR for the backend WVD VMs if when I switch over it will be in a different RG and Availability Zone/Region? How does the Broker and Web services know that the VM has been moved to a different place and now it has to redirect user requests?
From my experience, When i have done my DR testing the VM connects back to the WVD services via its agent as this is what tells both the again and the WVD what pool it is located in. theoretically i could look to deploy a VM in another region and join it to that pool in theory.
As the WVD agent checks in with the WVD Service to say its available in the form of a heartbeat.