Abstract
Context DevOps is becoming the de facto software development practice. As a DevOps teams releases new updates in days or even hours into the production environment, it is important to monitor the performance of the teams along the product lifecycle, for all the stakeholders, such as developers, managers, investors and customers, to have a clear and real-time understanding of the team productivity and the product quality. Traditional software engineering metrics, including DevOps metrics, such as cycle time, defect escape rate or developer satisfaction, take a long time to collect, or require document review, and do not allow for continuous monitoring. A recent trend for monitoring real-time systems is Digital twins, to enable simulation, modelling and optimization of physical entities. So far, no digital twin for monitoring of DevOps practices in modern software engineering has been realized. We want to use the digital twin as an approach to generate new insights for DevOps teams. Goal This M.Sc. thesis aims to identify how we can use a Digital Twin approach to continuously gain insight into the DevOps practices of software development teams. Method We have used design science research to develop an artifact, which includes 19 visualizations for 8 selected metrics. We evaluated this artifact with two case studies, using two open-source projects hosted on GitHub. Results The results show the insights that can be generated through the digital twin, as well as avenues for future work. Conclusion The findings support that based on the rich data collected by DevOps environments and tools, it is possible to build a digital twin of for the DevOps project and process for continuous data-drive monitoring of the performance of DevOps teams. The visualizations based on the digital twin can generate valuable insights into product stability, system and team performance.