I wrote the original Query Store performance overhead post just over two years ago, and just like the data in your database keeps changing, so does SQL Server. However, the question, “What is the performance overhead of enabling Query Store?” is still the most frequent question I am asked.
So why am I writing this post? Because there have been many improvements specific to Query Store that have taken the feature to the point where it can support all workloads, including those that are ad-hoc. This is a big deal.
Quick history (in case you didn’t read the initial post)