Which Azure Database Fits the Cosmos, 5-TB, and 30-TB Projects?

Why the Answer Is Correct

Project1 is a fast-iterating solution that consumes several Azure SaaS services, so the store must be schema-flexible, multi-model and globally distributable — that is Azure Cosmos DB, which exposes SQL, MongoDB, Cassandra, Gremlin and Table APIs and lets the data model change as the solution iterates. Project2 migrates a 5-TB on-premises database: Azure SQL Managed Instance accepts up to 16 TB per instance and offers near-100% SQL Server surface area (cross-database queries, SQL Agent, linked servers, Service Broker), which keeps the migration a lift-and-shift rather than a rewrite. Project3 is the same migration type at 30 TB, which is beyond the Managed Instance ceiling, so the only managed relational PaaS target that fits is Azure SQL Database with the Hyperscale service tier (up to 100 TB, fast snapshot-based scale and restore). The three selections therefore hinge on workload character for Project1 and on hard storage limits for Project2 and Project3.

Why Other Candidate Answers Are Wrong

Sending Project3 to Azure SQL Managed Instance is the most common error — a 30-TB database cannot be restored into an instance capped at 16 TB, no matter how SQL Server-compatible it is. Recommending plain Azure SQL Database (General Purpose or Business Critical) also fails on size: those tiers cap at 4 TB per database, so even the 5-TB project does not fit, and only Hyperscale breaks that ceiling. Azure Cosmos DB is wrong for Project2 and Project3 because it is a non-relational, NoSQL engine with no SQL Server compatibility — the existing schemas, stored procedures and Agent jobs would have to be rewritten rather than migrated. SQL Server on Azure Virtual Machines could physically host 30 TB, but it is IaaS: patching, high availability and backups remain the customer's responsibility, so it is not the recommended managed target when Hyperscale satisfies the requirement.

Community Comment Notes

Learners split on the two migration slots. One commenter (mesloth) maps Project2 to Azure SQL Database on the "5TB" keyword and Project3 to Azure SQL Managed Instance, while gpais answers the opposite way and points out that the "30-TB database exceeds the 16 TB" limit. gpais's size reasoning matches the documented Managed Instance resource limits, so the pair used here is Managed Instance for the 5-TB project and SQL Database (Hyperscale) for the 30-TB project. All commenters agree on Azure Cosmos DB for Project1 because it "integrates multiple SAAS" sources cheaply and read/write flexibly.