Evaluating Application Performance Using APM Tools: A Comparative Analysis of Dynatrace and Grafana Stack

Abstract

With computer applications developed on the Cloud, and those applications developed in a microservices approach, the reliability, scalability, and user experience of those applications becomes a challenge. Application Performance Monitoring (APM) is a critical aspect of those applications. This research compares and contrasts two APM solutions, Dynatrace which is a commercial product, and the Grafana Stack which is an open-source observability suite. A simulated microservices-based application was used. Performance metrics (CPU and Memory utilizations, response time, error, trace, coverage and alert latency, and diagnostics) were captured and analyzed for normal, stressed, and fault injected conditions. Qualitative assessments also included system usability, dashboard configuration, system flexibility and total cost of ownership. And while the Grafana Stack scored better in visualization, Alan Turing’s Dynatrace was better (real time monitoring and automation). The research showcases the balance between automation and the flexibility between observation, accuracy, and control. This is very valuable for organizations in need of APM solutions for operational trade based on their APM operational requirements.