The Battle for the Frontend: Rethinking APM

With the battle for the backend between data centers and cloud largely over, application development is the next frontier ripe for innovation. Studies project mobile app revenue at $935.2 billion by 2023 and Internet of Things (IoT) installations at $75.44 billion by 2025. It’s no surprise that software innovation is ramping up around application development in a big way. At the same time, the desktop is still a primary interface in the workplace, which has given rise to a wave of browser-based applications for enterprise collaboration and efficiency.

Apps—both on devices and in-browser—are effectively what businesses run on. New JavaScript-based Single-Page Applications (SPAs) are replacing legacy cloud stacks, where the software runs primarily in a customer’s browser. However, this approach creates an observability gap for teams under pressure to deliver better software more quickly. Because no application is siloed, the distributed nature of modern applications adds further complexity—and increases the risk for defects—for developers, who must code faster, smarter and better to keep their businesses competitive in an application-centric world.

The Battle for the Frontend

IDC predicts that by 2022, more than 60% of global GDP will be digitized, with growth in every industry driven by digitally enhanced offerings, operations and relationships—and applications will be the primary method of delivery. Companies in all industries are battling for users, revenue and reputation only, but only those companies with the best software will remain contenders. To innovate continuously, ensure optimal user experiences, and ultimately, retain their competitive advantage, enterprises must ship code quickly, every day and with fewer errors.

Application errors and crashes will dramatically spike churn rates, kill conversions and impact brand reputation and customer retention. As tolerance for faulty performance is rapidly diminishing, user experience becomes a business-critical concern that has a substantial impact on a company’s bottom line. While companies with faulty code are dealing with the fallout, those that rely on modern application performance monitoring (APM) are shipping better code more quickly and with less risk.

These new business drivers, along with an explosion of devices—from smartphones to IoT devices—have led developers to move their code closer to customers to increase code speed and improve scale. But this shift means software is run on customer devices over which developers have no control, a reality that inevitably creates its own set of new problems to monitor. To win this battle for the frontend, development teams that create games, mobile applications, IoT and other high-performance applications need rich insights into application health in order to quickly and continuously fix software errors, all with minimal impact.

Rethinking APM

Unfortunately, many legacy application monitoring tools do not solve for the frontend. They are fragmented, or specialized, and missing support for JavaScript, mobile, desktop apps or embedded IoT devices—despite these platforms being a significant part of the application ecosystem. Legacy tools focus more on system behavior, offering performance metrics on availability, throughput and latency, but miss the mark when it comes to actionable insight that helps frontend developers make sense of complexity and quickly get to the root cause of errors.

Modern applications require a holistic approach to monitoring that provides deep context and understanding about how to address issues when they arise, instead of simply displaying the symptoms of a problem. This approach should involve collecting and analyzing transactional traces, scaling search and storage to handle full traces, and building intelligent algorithms to detect and analyze issues. By moving monitoring from production environments to test environments, development teams can accelerate the software development lifecycle and reduce the impact on customers. APM must integrate with continuous delivery tools to monitor continuous health mechanisms.

— David Cramer