Real-Time Anomaly Detection with AI-Driven Syslog Monitoring for Bioinformatics Reliability
Sambasiva Rao Madamanchi · International Journal of Innovations in Science Engineering and Management. · 2025
In this study, we present a novel framework that integrates AI-based anomaly detection with syslog monitoring to enhance the reliability of microbial pipelines. Syslog data, typically underutilized in bioinformatics, is collected and parsed to extract features such as error frequency, runtime patterns, and resource usage indicators. We apply machine learning models including Isolation Forests and LSTM Autoencoders to identify deviations from normal system behavior in real time. Experimental evaluations on both real and simulated microbial workflows demonstrate high accuracy in detecting anomalies, including those that do not trigger pipeline-level errors. A key use case illustrates how the system prevents corrupted outputs caused by unnoticed memory faults during taxonomic classification.