Validate performance before/after index change
Posted on 2014-03-26
I was performing a database assessment recently, and I identified MANY duplicate and unused indexes. All of which, of course, we know cannot be too kind to application performance.
So, I structured the cleanup, to remove the dupes and unused indexes. But, before it can be run in production, I need to prove why this is even necessary, or why it should be done. I've been asked to run the same thing in staging to prove its worth, before doing so in production.
Basically, what is the most effective way of showing a before and after, or something that would help me substantiate that yes, duplicate indexes are an impediment, and should be removed.
Unfortunately, I am having problems coming up with the best way to validate this effort.