Benchmarking Best Practices in Application Performance Management

Application development and operations teams as well as database system administrators are charged with maintaining the uptime of both internal and external IT systems in their organizations. They are also charged with optimizing the often-numerous software applications running on these systems.

It’s a formidable task, by any measure, made even more formidable by the fact that systems and applications have evolved rapidly in recent years in both complexity and scale to now include public clouds, third-party services and an ever-growing array of advanced technologies. Today’s sophisticated, multi-tiered application architectures present businesses and end users with new opportunities to deliver value and improve the user experience but IT organizations with unprecedented challenges.

With the evolution in complexity and scale comes the need for next-generation application performance management (APM) solutions. These solutions have also evolved at a rapid clip and have become increasingly powerful and proficient at being able to identify and diagnose problems when critical software fails to work properly or operate at an optimal level.

The need is obvious when one considers that outages, sluggish response times and even glitches in both public-facing and internal applications can lead to dire consequences. Users suffer headaches and companies suffer financial losses — and, in some cases, irreparable damage.

Today’s most advanced APM solutions help organizations avoid nightmare scenarios by providing a highly automated, sophisticated and comprehensive approach to identifying application performance issues, no matter where they may occur, at the earliest possible point of detection. By helping to ensure that issues get detected and resolved in a time-sensitive manner, these next-generation solutions make it possible to avert, or at least minimize, the impact of a malfunction while at the same time instilling in the organization a culture of continuous performance improvement.

Over time, APM has evolved into a category of solutions within what is collectively referred to as a full-stack APM suite. This suite is comprised of four main components:

  • Web performance management (or user experience management)
  • Code-centric performance management
  • Operations-centric performance management

Some APM solutions are designed to enable code optimization. This category of solutions is used primarily by application developers who focus their day-to-day efforts on optimizing code, without paying much attention to possible issues that may reside with the underlying infrastructure. Utilizing APM solutions to monitor code, including looking for patterns of performance issues as users transact and interact with an application over time, developers can determine what parts of the code may have operational shortcomings and need to be addressed and within what timeframe.

Of course, many operational problems are not buried somewhere in a million lines of code, but, rather, in the underlying infrastructure. Some APM solutions are designed for use by IT professionals who are tasked with optimizing the underlying infrastructure resources and applications services. Similarly, some APM solutions are designed for use by database administrators who spend the majority of their waking hours working to monitor and optimize database operations. These two categories of APM solutions are infused with the context of storage devices, server hardware, relational databases, virtual machines and other elements and contingencies that come into play when it comes to ensuring that applications function at full capacity.

When application issues arise in a production environment, it is often because a service has stopped working, a database has slowed down, a virtual server is low on capacity, a dependent service is unresponsive or a resource contention has reared its bloody head. APM solutions geared toward diagnosing and troubleshooting specific areas of functionality may be the only way to identify the precise problem. These are certainly the only options for doing so in a time-sensitive manner.

In short, there are different categories of tools, all of which fall under the APM umbrella, for different types of IT professionals, based on their specific areas of accountability. The right tool for the job depends in large part on whether the professionals are part of an operations team or a development team. Do they need to monitor the performance of custom applications, built internally and for which they have access to the code, or commercial applications from external solution providers? Are they responsible for code profiling or operational troubleshooting?

Combined, and across all categories, next-generation APM solutions are a far cry from previous generations of monitoring tools, many of which remain in wide use, even in major enterprises. Beyond simple up/down monitoring, the older tools can execute queries that, if the organization is lucky, may hint at a problem before it escalates. However, the tools are generally ineffective in application response tracking, distributed transaction traces, and in providing other platform capabilities that make it possible to understand application operations and how issues like slow database or API calls are impacting performance, let alone pinpoint the specific source of a problem and identify the root cause in a timely manner.

The most advanced APM solutions, by contrast, excel in early detection, allowing organizations to discover potential issues before they become real issues. By providing a real-time window into application infrastructure, including actual code, a next-generation APM solution can reveal insights into the root causes of performance issues in much the same way that an MRI can reveal detailed pictures of soft tissues like the brain to detect tumors, traumatic injury, developmental anomalies, infection, and the causes of a headache.

Ultimately, organizations need to monitor for application-level issues as well as database- and infrastructure-related issues, any of which can throw a wrench into application performance. Focusing solely on database anomalies is a myopic approach and may well be a recipe for disaster; IT organizations need to have a comprehensive view into all aspects of application performance.

According to research conducted by Starfleet Research for The 2022 Benchmark Report on Application Performance Management, next-generation APM solutions enable application support teams, including developers and operators as well as database administrators, to improve application performance while identifying and diagnosing issues much faster and more accurately than in the past. In fact, 75 percent of IT organizations that have implemented an APM solution within the past three years have improved either “significantly” or “dramatically” the average application uptime and response times under peak loads.

More than two-thirds (69 percent) of IT organizations that have implemented an APM solution within the past three years have also improved either “significantly” or “dramatically” the average time required to identify and diagnose an application performance problem and production time for new application development, testing and implementation.

Of course, timing is everything when it comes to problem resolution. In today’s always-on environment, the difference between one hour and two hours of downtime can make all the difference in the world. Even within a software production environment, a performance problem is often treated as an emergency incident to be diagnosed and resolved as quickly as possible, and for good reason.

Automation is key to pinpointing root cause problems (including source code issues) faster and more effectively than in the past. With previous generations of software, the IT staff would manually gather and analyze data logs in the hope of identifying and diagnosing the performance issue. Next-generation solutions automate the process — and on a massive scale, when necessary. Assessing single-node performance is no longer a viable option for many IT environments. A next-generation solution can scale from hundreds to thousands of instances, meeting demand instantaneously.

According to research conducted Starfleet Research for The 2022 Benchmark Report on Application Performance Management, more than half (53 percent) of IT organizations that have implemented an APM solution within the past three years have improved either “significantly” or “dramatically” both customer satisfaction and employee productivity. That should be music to any executive’s ear.