Cut idle database spend with new Hyperscale Serverless controls

Azure SQL Database Hyperscale Serverless now pauses and resumes automatically. The new auto-pause and auto-resume feature is in public preview.

This is one of the clearest FinOps updates in the Azure batch. If a database sits idle for periods of time, compute can shut down automatically and come back when work resumes, which helps reduce spend without changing application behavior.

Give PostgreSQL and SQL Server more right-sizing options

Azure added new database and storage tiers for demanding workloads. Azure Database for PostgreSQL Flexible Server now supports Ultra Disk in public preview for I/O-intensive workloads.

That gives teams another performance tier to work with when storage needs are driving cost or bottlenecks. Having more options makes it easier to match storage spend to the workload instead of defaulting to a larger setup than needed.

Place compute more automatically across zones

Azure Virtual Machine Scale Sets now have automatic zone placement. In public preview, Azure can choose zones based on SKU availability and capacity.

That reduces the manual work of placing workloads across zones and can improve deployment efficiency. For teams managing lots of scale set capacity, it’s a helpful way to simplify operations while still using available infrastructure well.

Add vector search directly inside SQL Database

Azure SQL Database now supports vector search and vector indexes generally available. That means AI-style retrieval can happen directly in the database.

This can be a practical FinOps win because it may reduce the need for separate systems and duplicated infrastructure. If teams can keep retrieval inside SQL Database, they may cut down on extra tooling, extra storage, and extra operational overhead.