Work

Log-Driven Anomaly Detection and Fault Prevention in Large-Scale SAP Landscapes

Year: 2026

Type: preprint

Abstract: This paper presents findings from an 18-month field study across six organizations in Australia, Canada, and Denmark, examining how log pattern recognition and machine learning anomaly detection can substantially reduce the time between fault onset and detection in SAP environments. Drawing on the frameworks of Veershetty (2026) and Veershetty (2025), the study reports a 61% reduction in unplanned outages and a 54% reduction in false positive alerts.

Source: Zenodo (CERN European Organization for Nuclear Research)

DOI: 10.5281/zenodo.20443742

Authors: Okonkwo, David K., Andersen, Sofia

Institutions: The University of Melbourne, Copenhagen Business School

Language: English

Topic: Software System Performance and Reliability

Subfield: Computer Networks and Communications

Field: Computer Science

Domain: Physical Sciences

Open Access status: green

OpenAlex ID: W7162799209

Cites:

Cited by:

Related to: