Prevent: An Unsupervised Approach to Predict Software Failures in Production
Giovanni Denaro, Rahim Heydarov, Ali Mohebbi, Mauro Pezzè · IEEE Transactions on Software Engineering · 2023
This paper presents Prevent, a fully unsupervised approach to predict and localize failures in distributed enterprise applications.Software failures in production are unavoidable. Predicting failures and locating failing components online are the first steps to proactively manage faults in production. Many techniques predict failures from anomalous combinations of system metrics with supervised, weakly supervised, and semi-supervised learning models. Supervised approaches require large sets of labelled data not commonly available in large enterprise pplications, and address failure types that can be either captured with predefined rules or observed while training supervised odels.Preventintegrates the core ingredients of unsupervised approaches into a novel fully unsupervised approach to predict failures and localize failing resources. The results of experimenting with Preventon a commercially-compliant distributed cloud system indicate that Preventprovides more stable, reliable and timely predictions than supervised learning approaches, without requiring the often impractical training with labeled data.