In APMdigest’s recent article on Application Performance Management (APM) predictions for 2015, I gave my insight on the year:
“2015 will mark a significant shift in the way that APM tools are used by IT Operations teams. Driven by increased implementation of Hybrid-Cloud based applications and massively distributed applications, these teams will stop using APM tools as their “Go-to” primary tool – opting for unified Infrastructure / Application monitoring solutions instead. The APM tools will move into an integral code debugging solution for Developer-intensive DevOps processes.”
I wanted to elaborate on why this shift will accelerate in 2015. First, let me be clear about the importance of APM tools. They are absolutely necessary for developers and architects who do the deep debugging tasks when problems are actually in application code. The problem is that they don’t work well for IT Operations teams.
Application Management – Easy Enough in Three Tiers
In simple web applications, the “three-tiered app” is the standard. The middle tier – the application code/App Server – is the heart of the application. Most performance problems in simple web applications are caused by bad code, poorly designed integration, and misconfigured JVMs and CLRs. APM tools focus on the application code, and provide a powerful tool for finding and fixing production performance problems in these simple systems.
Application Complexity – Many-Tiered Apps Create New Problems
Virtual servers, cloud systems and creative architectural uses of these technologies have made the three-tiered application much less common. Applications built on top of these technologies are n-tiered, where n can truly be anything: 3, 5, 9, 17. As a result, the “Application Server” is simply one step of many rather than the central focal point. In this world, code problems are far less likely to be the culprit, and visibility at this tier doesn’t necessarily deliver good coverage of the full application.
Making things even more difficult, virtualization and cloud technologies are specifically designed to remove the direct “known connection” needed by APM tools for them to be useful outside the App Server, itself. In fact, these technologies make it difficult for any conventional management tool to see the application from end to end.
IT Operations – Business Service Delivery
While all these technology shifts have been occurring, IT Operations has been getting frustrated. They’re being asked to deliver more business services, serve more users and meet higher goals – all while managing a broader set of technologies with more complex configuration. And let’s not even get started on the shrinking budgets and reduced resources.
But the biggest issue for IT Operations is that while they’re responsible for end-to-end performance and availability of business services, their toolsets only manages pieces. Making matters worse, the “Go To” APM tool can’t even be used by 80% of the Ops team because it requires a working knowledge of code – not exactly in IT Operations’ wheelhouse. Even DevOps can’t fix this issue. At the end of the day, IT Operations requires THEIR OWN “Go To” tool that monitors business services, while helping solve problems no matter where they occur – without instigating a massive bridge call, or relying on precious subject experts like DBAs, Architects, Developers, etc.
IT Operations vs. APM (What’s Next?)
Ultimately, APM tools will continue to thrive, but rather than muddying the slippery slope of Dev doing Ops and “hoping” that they’ll find something in the code to fix, APM solutions will simply be the necessary deep dive tools for Developers and Architects to fix Application code issues.
Meanwhile, IT Operations is moving to more transactional solutions, not tied to code at all, that can see in the Data Center AND in the Cloud, and that actually help IT Ops teams solve problems quickly and easily.
I can’t wait to see what comes next, can you?
Vic Nyman is Co-Founder and COO of BlueStripe Software.