Smarter ITOps: Leveraging Generative AI for Metric Anomaly Detection and Incident Reporting

In modern Information Technology Operations (ITOps), timely anomaly detection, event correlation, and incident reporting are crucial for ensuring system reliability and minimizing downtime. This paper introduces a unified approach that integrates time-series forecasting with Generative Artificial Intelligence (Gen AI) to enhance the efficiency and responsiveness of ITOps. Our method employs Long Short-Term Memory (LSTM) models to forecast health metrics from software application containers, uses the Z-score technique for anomaly detection, and utilizes Gen AI for correlating anomaly events and generating incident reports. The proposed framework is applied to a dataset containing health metrics from a RabbitMQ container, which supports multiple dependent application containers. Key metrics include CPU, memory, disk, and network utilization. The experimental results validate the effectiveness of this approach in proactively identifying anomalies, uncovering correlations across events, and automating the incident reporting process. By combining time-series forecasting with Gen AI, this solution paves the way for a transformative shift in ITOps—enabling more proactive, intelligent, and self-healing systems. This integration not only improves operational efficiency but also enhances the experience for both organizations and end-users, marking a significant step forward in intelligent IT management. Key Words: ITOps, AIOps, Generative AI, LSTM, Metric Anomaly Detection.