In the world of performance monitoring, there a number of ways to identify problems. Each method is different and usually, neither way is better than another.
Over the years, I’ve started look at metrics around non-clustered indexes to help find problems. Once while working at a client site, I was astounded by how they would implement changes so rapidly. This quick rate of change usually had detrimental effects on the performance of the application. Of course as usual, once the change was in Production, it was difficult to fix things.