Improving La Redoute's CI/CD Pipeline and DevOps Processes by Applying Machine Learning Techniques
Ana Filipa Nogueira, José Ribeiro, Mário Alberto Zenha-Rela, Antoine Craske · 2018
The complexity inherent to software development and maintenance - not only in technical terms, but also from a human perspective - entails challenges that can be addressed as learning problems. Machine learning techniques may be employed as tools to gain insight about strategies that can lead to the improvement of the quality of software processes and products. Defect proneness prediction, in particular, may be identified as an active research field. As stated by DevOps guidelines, the possibility of obtaining quick feedback allows teams to operate in an agile mode in which communication, decision taking and problem solving are expeditious, allowing companies to boost business value. This paper describes ongoing research for applying machine learning techniques to improve the quality of processes and products inside the DevOps pipeline of the La Redoute's IT department.