Some solutions to poor query performance suggest creating extended statistics objects to help the query planner better understand the shape of the data. Because these statistics objects are not created by default, I assume that they have some non-neglible overhead.

What kinds of overhead are present, what statistics object parameters factor in to the overhead, and can I quantitatively measure what the overhead of a specific statistics object is?

Some possible kinds of overhead I've thought of:

  • Increased CPU usage
    • when inserting
    • when querying
    • when analyzing
  • Increased memory usage
  • Increased disk usage


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Browse other questions tagged or ask your own question.