Current trends in the field of APM (application performance monitoring)

A few years ago, most application processes were primarily monolithic and implemented on-premise. It was fairly easy to provide application monitoring by collecting selected metrics and records into logs, because their scope was manageable and all the infrastructure belonged to the given company or organization. Most applications were more or less independent of each other and their sole common factor was the data repository. This type of application (and organizational) independence enabled IT teams to choose any solution for application monitoring. At the time a unified, comprehensive solution that would cover an entire company or organization presented only minimal benefit.

Nowadays the application environment is being developed the opposite way entirely. Application/IT teams create services that are also used by other teams in the organization. Many processes were moved from the data center to the cloud, in which they used a variety of different cloud services. What's more, the number of key applications itself increased exponentially. An new requirement was thus posed – to completely (end-to-end) monitor performance across the borders of applications, right up to components the company or organization doesn’t own (e.g. cloud services and the applications/services of third parties).

Observability is a new standard in the field of application monitoring for native cloud architectures. It is based on an enormous quantity of collected telemetric data such as metrics, logs, events, and distributed transaction records, which measure the status or, if you will, the 'health' of application performance and behavior. And yet it is not the data collection that is most essential; its subsequent processing using artificial intelligence and a focus on significant deviations from normal operations is key. Algorithms thus make it possible to quickly determine the cause of an incident, that is, what causes a given application incident, and what specific behavior or event led to it. This helps developers understand not only what’s wrong in the system – what is slow or corrupted – but also how the problem was caused, where it occurred, and what effects it had.

With observability, IT departments, developers, and DevOps teams get not only an overview of applications but also other information related to their infrastructure, platforms, and the customer experiences these applications support and that depend on them. With this standard they can:

• find failures, software errors, unauthorized activity, and a decreased level of services provided
• keep informed about 'application health', i.e. system status, by measuring performance and resources
• understand how adjacent or dependent services can mutually influence each other
• find entirely new unknown states/incidents that have not occurred in the past
• identify long-term trends for capacity planning with regard to business goals and KPIs

The key capabilities of an APM solution that takes advantage of the possibilities of observability include:

1. Core driven by artificial intelligence – it automatically evaluates millions to billions of data points within a matter of seconds. With its help, IT teams can handle situations where they suddenly receive hundreds to thousands of alerts about possible incidents. Without artificial intelligence, it is nearly impossible for the teams to determine which alerts are relevant and which ones matter most.

2. Automated process – it monitors in real time all components in the environment in real time from the top down, including mutual relationships and dependencies, without the need for manual configurations or coding. With traditional monitoring tools metrics, protocols, routing, and user experience data are stored into data silos without context that would connect them and give them meaning. Automated, intelligent APM solutions supporting observability identify the user impact at the time of the incident in order that developers can focus on the most burning problems and quickly resolve them.

3. Distributed routing across open-source and native cloud architectures – it analyzes transactions in a complex way with the goal of eliminating gaps, blind spots, and incomplete records about the progress of the transactions.

IT teams operating multi-cloud ecosystems have the need for constant, automated, and AI-driven monitoring of the entire process and its individual steps and components. This is the only way for them to find, repair, and prevent application performance problems in real time.

The Dynatrace platform and its Davis® AI module automate cause analysis and reveal new unknowns even in the most complex cloud architectures. It ranks among the elite in the APM field, as shown by the fact that Gartner has now for the tenth time in a row declared Dynatrace a leader in the APM field and that Dynatrace has achieved 5 of the 6 highest scores in the Gartner report 'Critical Capabilities for Application Performance Monitoring (APM) report'.