When I deploy an instance through the portal, the structure of the FQDN my container is assigned is:
{domain label}.{seemingly random string}.{region}.azurecontainer.io
Deleting & redeploying an identical instance yields the same random string portion, so at least it's not randomized for each deployment, but as this does not conform to the structure defined here in the Azure docs, I'm left without any explanation. Is it a hash? An id? How can this be predicted before deployment?
I believe this is related to Prevent dangling DNS entries and avoid subdomain takeover and ACI DNS name reuse policy:
In order to avoid this, ACI will now allow customers to reuse DNS names while preventing DNS names from being reused by different customers. ACI secures DNS names by randomly generating a hash value to associate with the DNS name, making it difficult for another customer to accidentally create an ACI with the same name and get linked to the past customer's ACI information.
There are five "reuse levels" that use hashing to generate a unique DNS name for your container group:
noReuse - no hash, object's DNS name cannot be reused.unsecure - hash is based on the DNS name only (aka Any Reuse).tenantReuse (default) - hash is based on the DNS name and Tenant ID.subscriptionReuse - hash is based on the DNS name, the tenant ID and subscription ID.resourceGroupReuse - hash is based on the DNS name, the tenant ID, subscription ID and resource group name.