Start with APIs and CD: Rekindling Interest in APM - Part 1
The findings outlined in Part 1 of this blog point to a need for "smart" Application Performance Management (APM) solutions supporting automation of change monitoring, performance and availability management, and production troubleshooting functions. With such capabilities in place, Dev and Ops resources could be freed up to deliver the new software products that have become the lifeblood of the agile business.
Taken together, these findings make a strong case for APM investments. They also make a strong case for refocusing on APM as a research topic during 2016. I am currently in the process of developing a survey supporting a new research report to be delivered in mid-to-late June. Thanks to Apica, BMC and Riverbed for sponsoring the upcoming 2016 EMA APM research study entitled APM in the Digital Economy: What's Hot, What's Not and What's On the Horizon?
The study will investigate the types of capabilities IT organizations are seeking in a new generation of APM solutions as mobile applications, API-driven applications, containers, and streaming technologies hit mainstream. One important factor for IT organizations to keep in mind as they consider APM investments is the distinction between the monitoring and management functions, i.e. "Application Performance Monitoring" versus "Application Performance Management." The two differ primarily in two functional areas: depth of coverage and the sophistication of the analytics and other features supporting autonomy in problem detection and resolution.
The role of application monitoring is to quantify, correlate, store, and report granular metrics underlying end-to-end application/transaction execution. The management function goes a step further and begins to take over the expertise-driven analytical tasks traditionally done by human IT practitioners. Advanced APM solutions — in which the "M" stands for "Management" — add the analytics and deep-dive transaction visibility necessary to actually identify, at high levels of certainty, the actual root cause of performance or availability issues.
Features supporting the management function in today's leading-edge APM solutions include the ability to "learn" from environmental factors and use this learning to detect departures from normal behavior. Other features include predictive analytics supporting notification of impending issues/failures, along with autonomic capabilities supporting automated resolution (if a company chooses to take advantage of this functionality). However, perhaps the most important feature associated with these solutions is automation of the process of formulating insights from data and using those insights to draw accurate conclusions relating to "how do we fix it?"
While human experts have historically performed this function, automated products can do so far faster and more efficiently than their human counterparts.
From this perspective, investments in quality APM solutions can eliminate many of the production support tasks that are consuming the bandwidth of Dev and Ops teams. This, in turn, frees up these teams to deliver new software products and services at the fast pace necessary to achieve software-related business objectives.
Stay tuned for the research report, which is scheduled to be published in mid-June, and for the related webinar, currently scheduled for July 12.
