Anomaly detection in systems metrics

Leidinio https://doi.org/10.15388/DAMSS.12.2021The information technology (IT) sector is becoming an area where the highest value-added products are being developed. With the growing importance of IT systems, the public sector and business organizations are becoming responsible for data management and storage. More and more processes are managed by IT systems and the amount of data is constantly increasing, therefore security ensuring is becoming a challenge. Understanding the state of infrastructure and systems is essential for ensuring the reliability and stability of services. Information about the health and performance of the system helps to react to issues and make changes with confidence. One of the best ways to gain insight into the information of system status is to use a reliable monitoring system that collects metrics, visualizes data and alerts when things appear to be broken. In this research, metrics that represent the raw measurements of resource usage or behaviour that can be observed and collected throughout IT systems are analyzed. Anomaly detection in metrics is performed using ARIMA, TBATS, PROPHET, LSTM, and Isolation Forest methods, which can be applied to real-time operation to identify and report the detected anomaly as quickly as possibleBlue Bridge UABMatematikos ir statistikos katedraTaikomosios informatikos katedraVytauto Didžiojo universiteta