Blog Post

Azure SQL Blog
5 MIN READ

Azure SQL Managed Instance gets a boost: New capabilities in Next-gen General Purpose

UrosMilanovic's avatar
UrosMilanovic
Icon for Microsoft rankMicrosoft
Nov 17, 2026

This release of Next-gen GP closes the remaining major capability gaps with Classic General Purpose, while delivering new enhancements in availability, performance, and scalability.

When we introduced Next-gen General Purpose for Azure SQL Managed Instance, our goal was a faster General Purpose tier: by moving the data and log storage layer to Azure Elastic SAN, it enables higher performance, greater scale, and more flexible resource sizing.

Today, we are reaching an important milestone in that journey as this release of Next-gen GP closes the remaining major capability gaps with Classic General Purpose, while delivering new enhancements in availability, performance, and scalability.

What’s new in Next-gen General Purpose

The latest improvements span availability, performance, and scalability:

  • Zone redundancy is generally available, including support for flexible memory
  • Log throughput increases, both per vCore and at the instance level
  • More storage is available at lower vCore configurations

Each improvement brings meaningful customer value. Together, they mark something bigger: Next-gen GP is becoming the foundation on which we can take General Purpose forward.

Zone redundancy for higher availability and resiliency

Zone redundancy for Next-gen General Purpose is now generally available, allowing customers to improve application availability and protect production workloads against availability zone failures.

Zone-redundant instances can also take advantage of Next-gen GP resource flexibility, including configurable additional memory. Customers can independently right-size compute and memory while using a zone-redundant configuration.

Zone redundancy for Next-gen GP is available in all zone-redundant regions, except those where Azure Elastic SAN does not yet support zone redundancy. For the current list of supported regions, see the Zone redundancy section of the Azure SQL Managed Instance region availability documentation.

Zone redundancy availability is reflected in the Azure portal and updates automatically as new regions are onboarded. For more details on zone redundancy in Next-gen GP, see Zone-Redundant Next-Gen General Purpose for Azure SQL Managed Instance.

Higher throughput without requiring more compute

Next-gen General Purpose now delivers higher log throughput, both per vCore and at the instance level, enabling greater performance for transaction-intensive workloads.

The maximum instance log rate increases to 220 MiB/s, with improvements across Standard-series and Premium-series hardware:

  • Standard-series: Maximum log throughput increases from 120 MiB/s to 220 MiB/s, an improvement of more than 80%.
  • Premium-series: Maximum log throughput increases from 192 MiB/s to 220 MiB/s.

 

The diagram illustrates an increase in log rate (per vCore) with different vCore configurations, showing that the maximum log rate for both Standard-series and Premium-series reaches 220 MiB/s, with the largest configurations gaining the most performance improvement.

 

But the maximum is only part of the story - we are also increasing the log throughput available per vCore from 4.5 MiB/s to 6 MiB/s.  

On Premium-series hardware, the 40-vCore configuration now delivers up to 220 MiB/s, compared with 192 MiB/s previously. That is nearly 15% more maximum log throughput at the same compute size, while smaller configurations also benefit from the higher per-vCore limit.

The result is simple: more transaction throughput without requiring more compute.

More storage without overprovisioning compute

The same right-sizing principle now extends further to storage.

Historically, workloads whose storage requirements grew faster than their compute needs could be forced to move to a higher vCore configuration simply to unlock more storage.

With this update, storage scales much more closely with compute. The storage limit now rises with each step up in vCores: 4 vCores support up to 2 TB, 6 vCores up to 4 TB, and 8 vCores up to 8 TB. From 8 through 32 vCores, the ratio holds steady at up to 1 TB of storage for every vCore, as the chart below shows. As a result, you can get the storage you need without overprovisioning compute.

 

Chart showing how storage capacity scales with vCores: 4 vCores support up to 2 TB, 6 vCores up to 4 TB, and 8 vCores up to 8 TB. From 8 through 32 vCores, storage grows at up to 1 TB per vCore.

 

The practical impact is significant:

  • A workload requiring 12 TB can now run on 12 vCores instead of 16.
  • A workload requiring 20 TB can now run on 20 vCores instead of 32.
  • A workload requiring 24 TB can now run on 24 vCores instead of 32.

For storage-heavy workloads, this means better resource utilization and potentially better cost efficiency.

It also reinforces one of the fundamental ideas behind Next-gen General Purpose: compute, memory, storage, IOPS, and throughput should increasingly be resources you can size around your workload, rather than resources that force you to size one another.

Next-gen GP is becoming General Purpose

Next-gen General Purpose (GP) began as an evolution of the General Purpose architecture. By moving the data and log storage layer to Azure Elastic SAN, we created a foundation for greater scale, higher performance, and more flexible resource sizing.

That foundation has enabled:

  • Greater scale: up to 32 TB of storage, 500 databases per instance, and 128 vCores
  • Higher performance: higher IOPS and data throughput, faster storage scaling operations, and now up to 220 MiB/s of log throughput
  • More flexibility: configurable IOPS, flexible memory, and more granular compute choices

Performance and scale were only part of the journey. Before Next-gen GP could truly become General Purpose, it also needed to support the scenarios and capabilities that existing General Purpose customers depend on. With zone redundancy now available, the remaining major capability gaps with Classic General Purpose (GP) are closed.

Next-gen GP is now ready to take on the role Classic GP has played until now. Going forward, we plan to upgrade all Classic GP instances to Next-gen GP technology, consolidate on this new version of General Purpose, and eventually deprecate the Classic GP technology altogether.

What this means for you: you don’t have to wait for Microsoft to upgrade your Classic GP instances. You can get the benefits of Next-gen GP right away by upgrading them yourself. The upgrade is online, requires no application changes, and does not change pricing.

Stay tuned: we’ll soon share a more detailed plan, including Next-gen GP becoming the default and only General Purpose option, the service-orchestrated upgrade of Classic GP to Next-gen GP, and timelines.

Get started

The best way to experience Next-gen General Purpose is to try it yourself. With the Azure SQL Managed Instance free offer, you can try out Next-gen GP at no cost for the first 12 months, with 720 vCore hours of compute and 64 GB of storage included. You can create a free instance in the Azure portal, Azure PowerShell, Azure CLI, or the REST API.

As part of the free offer, you can choose the SQL Server 2025 update policy and try out the first SQL version built for the era of AI. Your instance stays aligned with the SQL Server 2025 database format, so you can use its engine features and keep the option to restore databases or replicate data back to SQL Server 2025.

Do you have an existing workload you want to try out in the cloud? The Managed Instance link makes it smooth, online, and easy. It uses near real-time replication to bring your SQL Server databases to Azure SQL Managed Instance, so you can test your workload on a read-only replica and cut over when you are ready, with minimal downtime.

Updated Nov 17, 2026
Version 1.0