Proactive Anomaly Detection Using Time Series Data of A Large Scale Platform

Abstract

Ensuring high availability of large scale e-commerce platforms and promptly detecting anomalies in system components is a challenging issue. In this paper, a novel end-to-end system that constantly tracks system components and clusters similar ones, in addition, detects and reports anomalies in the tracked system component has been proposed. This study has been tested on a real platform and its performance has been thoroughly analysed in terms of success and duration.