In this post, I show how I make decisions about unused or redundant indexes with the whole picture: analyzing activity across primary and all secondary replicas.