Using AI and machine learning for APM

Traditional APM relies on monitoring code execution to indicate problems, an approach that used to be enough for consistent application performance. However, modern apps typically consist of millions of lines of code often running in containers. Moreover, these application environments are interconnected and encompass both on-premises and multi-cloud environments. For example, research found that a single application transaction crosses an average of 35 different technology systems or components. To further complicate troubleshooting, IT teams must manage a broad spectrum of noncritical components that affect application performance as well as complex hybrid ecosystems that include Kubernetes orchestrations and innumerable containers.