Application Performance: Best Practices Do Work

This is a article discussing survey results collected by NetForecast, appearing in Business Communications Review, May 2007 - found here.  The basic theme is to survey how moving to APM Best Practices improved the overall IT management experience.  I was a bit confused by the use of "benchmarking" with APM, which I would call the APM Pilot.  But they do make a good definition of how process changes enhance the APM experience.

"The goal of APM best practices is to improve the performance outcome, and for the best outcome these best practices cannot stand alone. Each must be embedded into a continuous improvement process that ensures that application performance meets your business needs..."

They describe how "APM Benchmarking" will help an organization understand their capabilities and how they compare with other practitioners.  This is what I call the 'Skills Assessment' (Chapter 3), and with other exercises for assessing enterpise visibility, lets you understand overall maturity of your APM practice.  I created more of a idealized model, based on an amalgam of actual customer achievements.  But I never had the opportunity to count how many of each customers achieved a particular level of APM maturity.

What is interesting that among those companies that participated, here is where they 'self-assessed':

The bulk of the population rated themselves on the low side and this is consistent with what I have been seeing over the last few years.  Only 1 participant gave themselves a full rating, out of 329 total participants.  Folks really need best practices to help them progress in using APM to address performance problems and show value for the new processes that will be established.

There is also a nice chart that summarizes the types of metrics that the participants are interested in.

Again, the chart title doesn't make sense but these are the metrics that the participants are interested in.  The difference between what is important, and what is actually thresholded, confirms that visibility gaps exist and implementation maturity - which the authors define as "follow-through".  I prefer to use an APM technology stack, which implies the same metrics, but separating network, server and transaction monitoring - which tool will get you those metrics.

Another interesting graph is the impediments to realizing an APM practice.  THe authors call these "inhibitors" and it is a useful graph.

I explore these issues in Chapter 2 of my book s "entry points" and the situations which cause all of these varied ways to impede an APM initiative.