Making Sense of Failure Logs in an Industrial DevOps Environment

Processing and reviewing nightly test execution failure logs for large industrial systems is a tedious activity. Furthermore, multiple failures might share one root/common cause during sessions, the review therefore require redundant efforts. This paper presents LogGrouper approach automated grouping of to aid analysis enabling processing each log group as batch. uses state-of-art natural language clustering approaches achieve meaningful grouping. The evaluated in an setting both qualitative quantitative manner. Results show that produces good quality groupings terms our two evaluation metrics (Silhouette Coefficient Calinski-Harabasz Index) quality. shows experts perceive groups useful, are seen initial pointer root assignment.