Azure Sql Managed instance poorer performance than Azure VM or ANS server

Question

Azure Sql Managed instance poorer performance than Azure VM or ANS server

Mark Windle 40

We have a few public APIs running on an Azure VM with a separate Azure SQL VM. These having been running a fair few year, the servers are Windows Server 2016 and as part of modernising away from some other older servers hosted by ANS in Manchester, I have set up an SQL Managed instance.

I am finding the performance of the SQL Managed instance quite poor. Whilst one of the APIs may return in 3 seconds from the VM database and even 3.5 seconds if the database is hosted on the ANS server (I did this just for testing purposes), the Manage instance may take 5.5-6 seconds. This is pretty consistent in testing and I have no idea why.

Everything is on UK South, so while it may not be 100% optimised, I wouldn't expect it to be worse than a server in Manchester. Configurations are both servers are fairly similar with 4 cores, premium on the Managed instance, and also probably around 8 years newer.

I feel like I must have misconfigured something, so any advice would be appreciated. The same queries are being run on the same database backup copy, so should have the same indexes etc.

Sai Raghunadh M 3,885 Reputation points Microsoft External Staff Moderator

2025-04-25T18:24:10.87+00:00

Hi @ Mark Windle

Thanks for sharing the details — I can see you've done a thorough job comparing environments, and I can understand the concern when a modern service like Azure SQL Managed Instance doesn't perform as expected right out of the box.

In most cases, Managed Instance can offer excellent performance, especially when configured optimally. That said, some differences in architecture — like network isolation, query plan behavior, or I/O characteristics — can sometimes cause unexpected slowdowns, particularly if the same queries behaved differently on traditional VMs.

It might be worth checking a few things:

Query execution plans — even if the data and indexes are the same, slight changes in the optimizer due to a newer SQL engine version or compatibility level can affect performance.

Connection settings — Managed Instance defaults to encrypted connections, which may add a bit of overhead compared to your VM setup.

Wait statistics and I/O throughput — these can help identify if the instance is being throttled or waiting on storage performance.

Even though everything is in UK South, there could be networking differences between how Azure SQL Managed Instance handles connections versus a VM-based SQL Server.

Please go through this Documentation that might help you:

https://azure.microsoft.com/en-us/blog/key-causes-of-performance-differences-between-sql-managed-instance-and-sql-server/

The good thing is that these are often tunable, and Microsoft offers strong tooling (like Query Store, Performance Insight, and built-in metrics) to help pinpoint where the gap is.

Hope this helps. Do let us know if you have any further queries.
Sai Raghunadh M 3,885 Reputation points Microsoft External Staff Moderator

2025-04-28T16:53:50.9633333+00:00

Hi @ Mark Windle

We haven’t heard from you on the last response and was just checking back to see if you have a resolution yet. In case if you have any resolution

please do share that same with the community as it can be helpful to others. Otherwise, will respond with more details and we will try to help.
Mark Windle 40 Reputation points

2025-04-28T16:56:36.6+00:00

Hi. So it appears it was is mainly latency issue. Moving the API to a new VM within the Managed instances VNET seems to have brought them mostly inline.

Accepted answer

0 additional answers

Your answer

Sai Raghunadh M 3,885 Reputation points Microsoft External Staff Moderator

2025-04-25T18:24:10.87+00:00

Hi @ Mark Windle

Thanks for sharing the details — I can see you've done a thorough job comparing environments, and I can understand the concern when a modern service like Azure SQL Managed Instance doesn't perform as expected right out of the box.

In most cases, Managed Instance can offer excellent performance, especially when configured optimally. That said, some differences in architecture — like network isolation, query plan behavior, or I/O characteristics — can sometimes cause unexpected slowdowns, particularly if the same queries behaved differently on traditional VMs.

It might be worth checking a few things:

Query execution plans — even if the data and indexes are the same, slight changes in the optimizer due to a newer SQL engine version or compatibility level can affect performance.

Connection settings — Managed Instance defaults to encrypted connections, which may add a bit of overhead compared to your VM setup.

Wait statistics and I/O throughput — these can help identify if the instance is being throttled or waiting on storage performance.

Even though everything is in UK South, there could be networking differences between how Azure SQL Managed Instance handles connections versus a VM-based SQL Server.

Please go through this Documentation that might help you:

https://azure.microsoft.com/en-us/blog/key-causes-of-performance-differences-between-sql-managed-instance-and-sql-server/

The good thing is that these are often tunable, and Microsoft offers strong tooling (like Query Store, Performance Insight, and built-in metrics) to help pinpoint where the gap is.

Hope this helps. Do let us know if you have any further queries.
Sai Raghunadh M 3,885 Reputation points Microsoft External Staff Moderator

2025-04-28T16:53:50.9633333+00:00

Hi @ Mark Windle

We haven’t heard from you on the last response and was just checking back to see if you have a resolution yet. In case if you have any resolution

please do share that same with the community as it can be helpful to others. Otherwise, will respond with more details and we will try to help.
Mark Windle 40 Reputation points

2025-04-28T16:56:36.6+00:00

Hi. So it appears it was is mainly latency issue. Moving the API to a new VM within the Managed instances VNET seems to have brought them mostly inline.

Answer 1

Hi @ Mark Windle

I'm glad that you were able to resolve your issue and thank you for posting your solution so that others experiencing the same thing can easily reference this! Since the Microsoft Q&A community has a policy that "The question author cannot accept their own answer. They can only accept answers by others ", I'll repost your solution in case you'd like to "Accept " the answer.

Issue:

We have a few public APIs running on an Azure VM with a separate Azure SQL VM. These having been running a fair few year, the servers are Windows Server 2016 and as part of modernising away from some other older servers hosted by ANS in Manchester, I have set up an SQL Managed instance.

I am finding the performance of the SQL Managed instance quite poor. Whilst one of the APIs may return in 3 seconds from the VM database and even 3.5 seconds if the database is hosted on the ANS server (I did this just for testing purposes), the Manage instance may take 5.5-6 seconds. This is pretty consistent in testing and I have no idea why.

Everything is on UK South, so while it may not be 100% optimised, I wouldn't expect it to be worse than a server in Manchester. Configurations are both servers are fairly similar with 4 cores, premium on the Managed instance, and also probably around 8 years newer.

I feel like I must have misconfigured something, so any advice would be appreciated. The same queries are being run on the same database backup copy, so should have the same indexes etc.

Solution:

So it appears it was is mainly latency issue. Moving the API to a new VM within the Managed instances VNET seems to have brought them mostly inline.

Please remember to "Accept Answer" if any answer/reply helped, so that others in the community facing similar issues can easily find the solution.

Share via

Azure Sql Managed instance poorer performance than Azure VM or ANS server

0 additional answers

Your answer