How important is APM to your organization? Assuming you’re reading this in the 21st century (if not, come back to the future!), the answer is probably pretty important. Your applications are the business. The digital experiences of your customers depend on them, as does the productivity of your employees. And, according to a Gartner Survey on End-User Experience Monitoring, companies with between 1,000 and 10,000 employees have an average of 319 applications, while larger companies average almost 900; 81% of survey respondents cite end-user experience as one of the top reasons for purchasing APM solutions.
These applications are also quite complex, running on diverse platforms and consuming multiple services – some of which exist in the cloud. APM is necessary to help you understand and manage performance as experienced by your users, to understand the business impact of poor performance, to diagnose and correct problems, even to predict problems before they happen. No longer is APM simply an expensive insurance policy; no longer can you afford to monitor only your infrastructure and hope that your users are happy. Disgruntled customers impact loyalty, revenue and brand image, while unhappy employees lower productivity, reduce user satisfaction and damage the reputation of IT.
To varying degrees, modern APM solutions meet these challenges. And those of you who use Dynatrace application monitoring enjoy gap-free data and PurePath Technology® insights from the industry’s leading Digital Performance Management (DPM) platform.
Sounds fantastic, right? In fact, it is – with one significant caveat. Agent-based APM solutions focus on a core set of technology platforms, emphasizing Java, .NET, and PHP, and aggressively supporting the “new stack” architectures of the future. That focus, however, leaves significant portions of enterprise application portfolios in the dark. For most organizations, “mission-critical” applications of record rely on many platforms that are not supported by today’s agent-based APM solutions, or offer only a small subset of low-level capabilities. Prime examples include SAP, Oracle E-Business Suite, even Microsoft Exchange.
IT Deserves Better: Enterprise Visibility
IT operations teams should have insight into the performance and service quality of all their applications – regardless of platform. They should not need to rely on app teams to uncover problems, nor should they be expected to adopt a series of disparate and complex platform-specific tools. In fact, many of these tools are designed to analyze system component performance rather than the application as a whole, and rarely include effective insight into end-user experience.
These realities have left many IT teams continuing to support business-critical applications with infrastructure performance monitoring (IPM) solutions, an increasingly discredited approach to understanding service quality. (I won’t go into any detail here about the shortcomings of IPM for monitoring service quality, but if you’re interested, there’s a short eBook with some additional insight.)
Don’t Get Fooled Again
AA NPM solutions teased at addressing this gap by adding “application awareness” to network probe-based monitoring solutions. However, the lightweight definition of “application awareness” and the dumbing down of the term “end-user experience” generally relegated these solutions to the sole use of the network team, without delivering on the promise of effective collaboration with business or development peers.
Let’s take a fresh look and define a set of criteria for what we’ll temporarily (at least for the duration of this blog) call “probe-based APM” – a platform-agnostic APM-like solution. The four high-level capabilities outlined here map loosely to a subset of Gartner’s five dimensions of APM. Obviously they exclude deep-dive visibility into application code, since that insight requires agents, but the reality is that IT teams do not always need – or want – code-level insights. Included in its place is deep-dive visibility into network performance, leveraging a network probe-based architecture.
These are of course rather high-level capabilities, supported by many underlying features. See if you agree with the list.
The Four Dimensions of Probe-Based APM
Capability 1: Measure the end-users’ experience
Today, end-user experience monitoring is considered core to performance management. But be careful; the definition of EUE should be “click to glass,” not a proxy metric inferred by network quality or application “responsiveness.” This is critical to provide context to infrastructure metrics and to help prioritize IT actions. It answers the fundamental question “should I care?” Without this measurement, how will you know if you have a problem that impacts users? Unless the problem is catastrophic, you won’t know until they call. You’ll also chase problems that don’t exist – behavioral “curiosities” that have no impact on EUE.
EUE measurements are also the fundamental measurement of service quality used to collaborate with business peers and drive business alignment, helping you understand the business impact of performance. Without these measurements, you have only “internal” metrics – resource utilization, availability, etc., – and risk permanent outsider status at the executive table.
Capability 2: Transaction performance visibility at each data center tier
By decoding and measuring transaction performance at each dependent tier, the solution can correlate degraded EUE with anomalous backend transaction performance, reliably isolating the fault domain. Without this visibility, you won’t be able to isolate the fault domain with any consistency; attempts to correlate EUE with server performance metrics are often misguided, leading to distracting red herrings.
Transaction names and their parameters are also the primary method used to collaborate with development peers and speed problem resolution, and capturing this context is important. Without it, you will not be able to provide the information necessary to hand off the problem to the development team. Instead, you’ll end up pointing fingers and losing credibility.
Capability 3: Measure the dynamic network’s impact on EUE
In addition to switched/routed network connectivity, modern corporate WANs are increasingly characterized by application-fluent devices such as Application Delivery Platforms (ADPs), WAN Optimization Controllers (WOCs), VPN tunnels, even Citrix XenApp servers; these sit in the path between remote users and data center applications. Measuring the performance impact of these complex application delivery chains is critical for comprehensive visibility into user experience. Without measuring what is often the most complex and problematic part of the end-to-end system, the “network” will often be blamed – with or without justification – for poor performance, resulting in wasted efforts, counter-productive tuning, and unnecessary delays in service restoration. And just as capturing transaction parameters to aid the app team’s analysis, on-demand or triggered packet captures are important to support deep-dive network analysis.
Capability 4: Monitor all enterprise applications – web and non-web alike.
You monitor the performance of your web-based apps of engagement, since they’re public-facing and revenue-driving. Why shouldn’t you monitor the applications your employees depend upon with similar consideration? They too have tangible impact on business success. Without insight into the user experience for enterprise applications of record, employee satisfaction and productivity suffer and business costs rise; ultimately, these factors indirectly, but tangibly, impact your customers. In fact, these apps of record often interface directly with your apps of engagement, resulting in a more direct link between their performance and customer impact.
Fear No More
Many of your enterprise applications of record are increasingly critical to the overall success of your business; their platform pedigree alone should not dictate the level of care or custody. To see how your peers use Dynatrace Data Center Real User Monitoring (DC RUM) for enterprise visibility into their broad application portfolios, visit our website. Interested in more? Join Kris Ziemianowicz, end-user experience expert and Senior Product Manager at Dynatrace, for a live webinar on why DC RUM solution should be a critical part of your APM strategy. Register now, or watch for additional sessions to accommodate your time zone preference.