Before we kick off the preview, we’d like to hear from customers interested in trying it. The new capability will let you replicate data from SQL Server to a managed instance that’s part of a failover group, combining hybrid replication with geo-redundancy in Azure.
The preview isn’t starting just yet. We expect to begin onboarding customers in the coming weeks, and capacity will be very limited. If this is something you’d like to try, let us know by filling out the sign-up form.
Register your interest in the upcoming limited preview: https://aka.ms/milink-fog-signup
What are the Managed Instance link and failover groups?
The Managed Instance link enables near real-time data replication between SQL Server and Azure SQL Managed Instance. Built on distributed availability groups, it replicates user databases in near real time from SQL Server hosted anywhere to Azure SQL Managed Instance, or from Azure SQL Managed Instance to SQL Server 2022 or later.
Customers use the Managed Instance link for:
- Offloading read-only workloads to Azure, or from Azure back to on-premises.
- Online migration from on-premises to Azure, or from Azure back to on-premises, with minimal downtime.
- Bidirectional disaster recovery between SQL Server and Azure.
Failover groups, also known as geo-replication, manage the replication and failover of user databases from one managed instance to another managed instance in a different Azure region. They provide read-write and read-only listener endpoints that remain unchanged after a geo-failover, so applications don’t need connection-string changes. The geo-secondary can also take on read-only workloads.
Until now, you couldn’t combine the two scenarios. Establishing a two-hop replication path from SQL Server to Managed Instance 1 and then to Managed Instance 2 presents a technical challenge and isn’t supported today. As a result, you’ve had to choose between using the MI link with a single managed instance or using failover groups without the MI link.
The capability we’re getting ready to preview removes that trade-off. SQL Server replicates to MI-1 through the Managed Instance link, while the failover group keeps MI-2 in sync in a second Azure region, as shown in the figure below.
Figure 1: SQL Server replicating to a primary managed instance through the Managed Instance link, with the failover group geo-replicating to a secondary managed instance in another region.
Replication in both directions
In this configuration, the link works bidirectionally. You can replicate from SQL Server to Azure and back from Azure to SQL Server. In practice, that means you can fail over to Azure and, when you’re ready, fail back to SQL Server.
Version requirements follow the existing MI link documentation. Two-way replication and failback require SQL Server 2022 or SQL Server 2025, with the managed instance on the matching SQL Server 2022 or SQL Server 2025 update policy. SQL Server 2016 through 2019 support only one-way replication to Azure.
What this unlocks
- Hybrid disaster recovery with regional resilience in Azure: Keep SQL Server as your primary, protected by Managed Instance in one Azure region and a geo-secondary Managed Instance in another Azure region.
- Migration to Azure with minimal downtime: Land directly in a geo-redundant configuration without having to re-seed or re-architect after cutover.
- Reverse migration: Move from a geo-redundant Azure configuration back to on-premises SQL Server.
- Read scale-out: Offload read-only workloads to a geo-redundant Azure SQL Managed Instance configuration.
- Reverse data replication: Replicate data from a geo-redundant Managed Instance configuration back to on-premises SQL Server.
- Who should apply and how the limited preview works
If you run SQL Server workloads that could benefit from both hybrid replication and geo redundancy on Azure SQL Managed Instance, we’d love to hear from you.
A few things to know
- The limited preview hasn’t started yet. Right now, we’re collecting interest from customers who would like to participate.
- We expect onboarding to begin in the coming weeks.
- Capacity will be limited, and the product team will contact selected participants directly.
- The preview isn’t intended for production workloads unless agreed with the product team.
- We’ll be looking for feedback from participating customers to help us shape the capability as it evolves.
Interested? Let us know
If combining the Managed Instance link with failover groups could simplify your hybrid, disaster recovery, or migration architecture, register your interest today.
We’re looking forward to hearing about your scenarios and working with some of you as we get the preview underway.
Register your interest in the upcoming limited preview: https://aka.ms/milink-fog-signup
Learn more
• Overview of the Managed Instance link
• Overview of Failover groups for Azure SQL Managed Instance