Data analytics for devops effectiveness

This paper discusses the opportunities and challenges associated with the data-driven approach to DevOps. The authors present analytical methods and techniques that can be applied to data collected from the DevOps process, as well as several ways in which that data can be used to improve the enterprises’ development capabilities. The authors include specific recommendations regarding the data that should be collected over time, as well as common data storage best practices for enabling analysis and reporting. Metrics and DevOps effectiveness KPIs are described in the paper as well. As an example of KPI analytics the authors show particular application of machine learning algorithms for classification of new code change requests into cost categories, which facilitates deployment activities optimization and cost reduction. The authors’ proposed approach is explained in the context of use cases at existing internationally recognized leading companies, which run multiple large-scale software development projects simultaneously. In addition, the paper explores the changing role of the DevOps engineer regarding data analytics, and highlights the requisite skills and knowledge for him to be successful in the big data era.