On AWS and GCP, a Databricks account is an independent entity, so having one Databricks account per environment (Dev/Test/Prod) is a straightforward approach and naturally results in separate Unity Catalog metastores per environment.

For Azure Databricks, I'm trying to understand the rules and limitations around metastores and Databricks accounts.

My understanding is:

• An Azure tenant can have only one Databricks account.
• A Databricks account typically has one Unity Catalog metastore per region.

However, during my testing and research, I came across references to this being a soft limit, and it appears that it may be possible to create multiple metastores in the same region within a single Databricks account. The main restriction seems to be that a workspace can only be assigned to one metastore at a time.

Based on that, I have a few questions:

1. Is it officially supported to have multiple Unity Catalog metastores in the same Azure region under a single Databricks account?
2. Can a single Azure Entra ID tenant have multiple Databricks accounts, or is the relationship strictly one Databricks account per tenant?
3. If separate Dev and Prod metastores are required for stronger isolation, what is the recommended Azure Databricks architecture to achieve this?
4. Would this require:
   • Multiple metastores within the same Databricks account?
   • Multiple Databricks accounts?
   • Multiple Entra ID tenants?

I'm looking for guidance based on official Azure Databricks/Databricks recommendations and supported architecture patterns.