If you are only monitoring your application’s performance, then you are missing 60% of the problems in your infrastructure. Gartner Group found that application related issues account for only 40% of infrastructure failures. The other 60% of failures are composed of technology and operator error. [http://goo.gl/q1qYDT]
To achieve 100% visibility of your infrastructure, you need to monitor more than just applications. As common as a bug in your application’s code is operator error. These types of errors are usually found in configurations of the various pieces of the infrastructure. Therefore one must monitor the entire infrastructure stack, from hardware to application.
Beyond operator error, databases, network, servers, and various hardware at all levels can be the cause of a failure, or worse an intermittent problem or slowness. How important is slowness to an application? Marissa Mayer, while working at Google, reported that, “an additional delay of 0.5 seconds to page-load time caused a 20% drop in traffic.” [http://goo.gl/iu6VWF]