Assessment of Real-World Incident Detection Through a Component-Based Online Log Anomaly Detection Pipeline Framework
Scott Lupton, Lena Yu, Hironori Washizaki, Nobukazu Yoshioka, Yoshiaki Fukazawa · 2023
This study introduces an open-source, component-based pipeline framework for online log anomaly detection. It implements popular parsing, encoding, and anomaly detection methods as replaceable components, and compares their performance to industry-standard, rule-based methods using real-world incident log data. The goal of this study is to assess the suitability of using modern log anomaly detection methods for industry system monitoring.