Every so often, Microsoft puts some extra into their SQL Server product, which doesn’t get as much time in the spotlight as it should. Nope, not talking about any of the great, new things that we got coming in with SQL Server 2022. Today’s post will be used to look at something called. SQL Assessment API. It’s not a very exciting name for a feature, but it definitely provides a new, exciting capability!
So, what exactly is the SQL Assessment API, and where, and how would you use it?
Not too long ago, Microsoft released The Final Service Pack for any version of SQL Server, as a part of their move to a Modern Servicing Model. This (and the fall time here in Finland) got me again thinking about the benefits of running your database workloads on PaaS over self-managed VMs. And one of the very real benefits is, that PaaS databases tend to be running with up-to-date versions. This is what you might have seen called as “evergreen” by Microsoft. When it comes to SQL Server running on VMs, well, these tend to be more of a “nevergreen” type of deployments. At best, they are somewhat yellow, but almost always never green.
And this isn’t an on-premises problem only, patching self-managed database services in the cloud is just as unpleasant as it is on-premises. But unlike death and taxes, continuously patching your operating system and the database software is something that you can avoid, without having to fear cosmic powers or a prison. It does require us to adjust our thinking on how we’re managing these systems, though: The databases aren’t the same as the virtual machines, and your virtual machines should be treated like they are disposable, not like pets.
I was just recently going through my backlog of podcasts and came across Data Exposed episode with Microsoft’s Pam Lahoud (@SQLGoddess on Twitter) about SQL Server VMs in Azure. Now, we have it from the official sources that one of the best VM sizes for SQL Server workloads is the Edsv4-series.
Incidentally, this is also one of my favorites for running SQL Server, and one that I often end up recommending due to missing monitoring data. Read further to find what exactly makes this VM series so good for many of the SQL Server workloads.