Press "Enter" to skip to content

Category: Cloud

Load Testing Azure SQL Hyperscale

Reitse Eskens cranks up the speed:

My previous post on load testing Azure SQL covered the Business Critical tier; let’s move to the one that is getting a lot of attention lately: Hyperscale. And no, you’re not the only one with Scooter’s Hyper Hyper song playing in the back of your mind.

Before we get into the tests, I need a moment to explain a few things about this tier. No, it’s not as complicated as the DTU models, but there are a few things to keep in mind.

Read on for those things to keep in mind, as well as how this performs compared to other options like Azure SQL Database Business Critical.

Leave a Comment

What’s New at Ignite 2026

Arun Ulag tells us what’s new:

Over the past year, my conversations with customers have changed. A year ago, the question was which model to choose. Now it’s how to put agents to work on the processes that run the business, and how to trust them to do it.

With the same frontier models available to everyone, the advantage comes from what no one else has: your context. It’s how your business runs, what your people know, and the systems and data you’ve spent years building. That’s your IQ. It’s your intellectual property, so it should stay yours, and it should work wherever your people and agents do. Our job is to make every layer beneath your agents better, from the infrastructure that runs them and the estate they act on to the apps they power and the business context that tells them how your company works.

If you’re all-in on agentic AI, you’ll find a lot to be happy about. If not, well…

Leave a Comment

Tracking How Many Row Versions an RCSI Query Reads

Louis Tochon works with a new bit of information:

The Table line looks like a normal scan of 330 pages. The Object line says the read walked through 220,000 row versions, 22 per row, because another session had updated these rows 22 times in a transaction it had not committed yet. The same read took about 117 ms of CPU, against 3 ms when nothing was pending.

The Object line is new. It comes from the version store read metrics Dimitri Furman announced on the Azure SQL blog last week, which also appear in the actual plan XML returned by the server (SSMS 22.12.0 drops them, both in the graphical plan and in Show Execution Plan XML; copy the XML cell from SET STATISTICS XML ON to see them), in four extended events, in sys.dm_exec_query_stats and its procedure, trigger and function counterparts, and in sys.dm_exec_requests and sys.dm_exec_sessions (Microsoft documentation). For now they exist in Azure SQL Database and SQL database in Microsoft Fabric; Managed Instance and the next SQL Server version are announced to follow.

When I recommend RCSI, the usual worry is the open transaction someone forgot in SSMS. These metrics let me check what it actually costs the readers.

The article looks to be AI-generated, but it does show off a new bit of functionality that I’d be happy to see spread beyond Azure SQL Database.

Leave a Comment

Changes in Microsoft’s Data Governance Strategy

Victoria Holt lays out some news:

AI is forcing a rethink of data governance. We are moving away from governance being something that sits alongside the data estate towards it becoming part of how data is managed and used.

The problems are familiar. Governance becomes fragmented across platforms and teams, controls become difficult to manage consistently, and manual processes struggle as the data estate grows. None of this comes from bad intentions. Data estates simply grow faster than the governance around them.

Putting governance closer to the data makes sense, particularly when that data is increasingly being used by AI as well as people. It makes it easier to use. An AI agent needs to know whether it can use the data, whether it is sensitive, what it means and whether it can be trusted. That makes good governance more important, not less.

Wolfgang Strasser has some additional thoughts:

Instead of separate Purview Data governance consumption meters, Microsoft states that governance is included in Fabric capacity. Actual cost implications still requires analysis. This is one of the things I am worrying about. For me, foundational governance features must be part of the platform and not be directly built using capacity units.

The cost is a question. The other question is, will Purview matter after this? Purview was a really neat idea, but I’ve always considered its cost prohibitive for most organizations. And, in one year, it’s going to be less capable but probably not less expensive.

Leave a Comment

Thoughts on Automatic Index Compaction

Monica Morehouse asks a question:

hile attending SQLCon and FabCon Europe in Barcelona, there were a lot of announcements that caught my attention. But there was one during the SQL keynote that made my DBA ears perk up: Automatic Index Compaction is generally available. And then came the comment that really got my attention: Could this be the beginning of the end of traditional index maintenance? Wait…WHAT? For those of us who have been working with SQL Server for a long time, index maintenance has practically been part of the DBA job description. Check fragmentation. Reorganize. Rebuild. Repeat. And for far too many environments, the strategy has basically been:

It’s Sunday at 2:00 AM. Time to rebuild ALL THE THINGS!

But should we really be doing that?

Click through for her answer. Also, shout out to Greg Moore in the comments for the Betteridge’s Law of Headlines usage. Greg is clearly a classy individual full of great ideas. Jeff Moden, additionally, has a comment worth reading.

Leave a Comment

Thoughts on the Databricks Lakehouse

Miles Cole writes from inside the company:

The lakehouse architecture introduced a simple but powerful idea: unify the robust data management of the data warehouse with the openness, flexibility, and decoupled storage of the data lake. In principle, a lakehouse can be implemented anywhere, on any object store, with any compute engine that supports open table formats. In this post I want to make the case for why I believe the lakehouse is best realized on the platform that invented the product category in the first place: Databricks.

I’ll make the case with three points:

  1. Robust data management comes standard
  2. Flexibility without the complexity
  3. A track record of innovation

Click through for the argument.

Leave a Comment

Hash Keys and Synapse Tables

Andy Brownsword explains the importance of selecting an appropriate hash key:

In Synapse SQL Pools, Hash distributions allow us to somewhat control data distribution. Whilst this can deliver performance in the right scenarios, I noted last week that these tables should be monitored to avoid skew.

In this post I want to demonstrate how to identify the skew, and why a different distribution model may be better suited.

One of the key tricks is that your choice of hash key should be approximately unique. There’s no explicit uniqueness constraint, but having a column with high cardinality and relatively little skew leads to a good distribution.

Leave a Comment

The Pain of Azure Defender Billing

Steve Jones walks through a painful experience:

I hadn’t done much in the last couple of months on Azure. For some reason, I decided to check my bill. What I found was that there were charges for $35-ish/month. This is within my free credit limit as an MVP, but that seemed high. I have a few serverless dbs and some storage, but I was expecting more of a $15 charge.

Azure billing is probably 85% of what an ideal billing solution would look like. There’s a lot of detail and you can get the bill down to a pretty granular level. But the problem is, like Steve found out here, that it’s really hard to tie back individual line items to specific services in a consistent way. Sometimes it’s because of ancillary costs like networking bandwidth. Sometimes, it’s a problem where you can’t easily see what specific resource is costing you money. Either way, it turns a proper audit of the bill into a bit of an ordeal.

Leave a Comment

Azure SQL DB Hyperscale versus Distributed SQL Solutions

Anna Hoffman clarifies:

In my last post, I broke down the Socrates paper, the research that became Azure SQL Database Hyperscale. Writing it made me want to tackle a question I think trips a lot of people up when they’re choosing a database: is Azure SQL Database Hyperscale “distributed SQL”? 

It’s a fair question. “Distributed SQL” has become shorthand for “the relational database that scales in the cloud,” and anything outside that bucket tends to get filed under “the relational one my company already has.” I’d rather not leave that to assumption, so here’s how I think about it. 

Read on for Anna’s take.

Leave a Comment

Load Testing Azure SQL GP and BC

Reitse Eskens continues a series on load testing. First up is Azure SQL Database General Purpose edition:

My previous post on load testing Azure SQL covered the Premium Edition; today I’m moving to the General Purpose Tier.

The GP tier comes in two flavours: regular and serverless. Serverless has only one hardware tier: Gen5. With provisioned, you can also select memory-optimised or confidential compute. In earlier tests, I never found significant differences between serverless and provisioned, so I skipped testing both SKUs. Given the nature of these tests, going memory-optimised didn’t seem to make much sense either. Also, I’ve never seen any other deployment than the regular Gen5 General Purpose database.

And then there’s Business Critical:

My previous post on load testing Azure SQL covered the General Purpose Edition, let’s move it up to the Business Critical Tier

The Business Critical tier comes in two flavours: General and confidential compute. For my testing, I’ve stuck with the General SKU. I haven’t seen confidential compute with customers, and it has very specific use cases: you want to protect your threads on the processor itself. A lot happens behind the scenes to make that work, and it costs more. There are scenarios where this makes sense, but not in my work. For example, using 8 cores for regular business-critical workloads costs about 3,650 euros each month. Confidential Compute will cost about 6,000 euros each month. So yeah, you need a use case for that.

Click through for results on each.

Comments closed