Shake Up Your APM

One of the challenges surrounding APM is figuring out what do do with our legacy management tools. Sure, end-to-end APM makes sense in a new, greenfield environment where we can drop in a just purchased suite of tools to provide holistic monitoring. There are many software tools that can help us accomplish this. However, the reality is that most IT organizations have already invested in an array of framework or point management solutions. The challenge for them is how to eek more performance out of tools that really can't deliver the needed solutions.

With costly enterprise network management systems providing packet loss, latency and other critical infrastructures metrics, system agents were then added on to try and get a more complete picture of the environment. Unfortunately, this presented silos of data that sometimes were all forwarded up to a manager of manager. We rarely were able to transform that data into information that could actually improve the customer experience. Then the software industry started to get smart and market business service management or IT service management. Unfortunately, we needed a new suite of tools for this too.  

While the BSM strategy resonated with the business side, it was met with a frustrated IT organization that still saw finger pointing between the network and application groups when performance problems arose. With the evolution and convergence of tenants of BSM and end-to-end APM, the drumbeat of focusing on the business user and performance is even stronger. For IT, the pressure to provide end-to-end metrics is intense.

For an enterprise architecture perspective, the challenge lies in the ability to provide a level of visibility and leverage an existing investment in other monitoring tools. That challenge is compounded by the migration of the enterprise application architecture to virtualization and cloud delivery models. I see two choices.

Choice 1: Comfortable with a suite of management tools and vendors, organizations are trying to retrofit legacy tools and provide end-to-end APM capabilities through the use of business analytics. Thinking of their enterprise management suite as metric collectors, they are abandoning much of the visualization and correlation layers of the architecture and replacing them with more robust business intelligence and analytics tools. While this leverages an existing investment, it often produces more complex solutions that still struggle with getting to the root cause of the performance issue and monitoring user experience.Choice 2: Organizations are evaluating their true business needs against APM enterprise architecture and starting from scratch. The technology advances around APM and the changes to the enterprise application architecture warrant a revisiting on the entire suite, and as painful as it may seem, changing to another end-to-end monitoring solution may save you in the long-run.  Swallowing that reality, however, is hard to do and only through leadership at the executive level will organizations have the courage to undertake this task. Politically and economically replacing products is hard, sparking fear of even longer deployments and more complex environments than the previous one.

Too many organizations get stuck with solutions that don't meet their needs. Everyone is afraid to stop the project or toss out the tool for fear of being ridiculed for making a bad decision. The safer solution is to continue to throw good money after bad and get trying to make incremental progress toward an ever more elusive goal. This is one of the reasons some business leaders are turning to cloud computing and doing and end-run around IT.

Evaluating the APM enterprise architecture is a good place to start if you are tangled in a quagmire of management tools that still do not meet your needs. You may be surprised with the options that you really have and the recognition from the business leadership for making a bold decision to take another look. Political suicide or pragmatic leader? I'd like to hear your comments.