The trigger for this feast were a number of questions from my employer. What database tier and SKU (Stock Keeping Unit) should we choose for customer C. Well, the answer depends on their budget and the performance they’re expecting. And we didn’t have a conclusive answer. Microsoft will provide you with their statistics on database performance, but what does that mean in real life. Because no matter what the good people in Redmond can come up with as databases, we as data engineers can do worse. Much worse.
So, I created some horrible query’s to insert, select and delete data and fired them off to a database. My own laptop with nvme disks provided the baseline, after that I gradually worked my way up from the Basic DTU database to the Hyperscale database.
This first post is an introduction to the method and laying out expectations. I’m looking forward to the remaining posts in the series.