Query Store is the answer to the challenges described previously. It was introduced in SQL Server 2016 and extended in SQL Server 2017. It collects the most relevant information about executed queries: query text, parameters, query optimization and compilation details, execution plans, execution statistics (execution time, CPU and memory usage, I/O execution details), and wait statistics; Query Store stores them in a database so that they are available after server restarts, failovers, or crashes.

You can use Query Store not only to identify performance issues, but also to fix some of them. Query Store offers a solution for issues caused by changed execution plans. By using Query Store, you can easily enforce an old plan; it is not required to rewrite the query or to write any code. You don't affect the business logic, therefore there is no need for testing; there is neither code deployment nor an application restart. By taking this approach, you can quickly implement a solution, or at least a work-around, and save time and money.

Query Store also captures some information that is not available in the server cache, such as unfinished queries or queries with broken execution, either using the caller or by an exception. You can easily find out how many executions of a single query were successful and how many executions ended with exceptions. This was not possible prior to SQL Server 2016 by querying the server cache.