An Anomaly Localization Method for Large-Scale Micro-services
Weiting Zhao, Bintai Xu, Chao Ma, Shaosong Zhu, Yu Tai, Gang Li · 2024
As the scale of micro-service systems on the cloud continues to expand, the dependencies between micro-services become more and more complex. Service anomalies may continue to propagate with the call relationship, leading to a decrease in the overall availability of the system. In order to cope with large-scale micro-service systems with complex dependencies, this paper designs an anomaly location method MicroALLS suitable for large-scale micro-services. This method collects micro-service operation indicator data based on Istio; uses the anomaly detection module of the variational auto-encoder for anomaly detection; once an anomaly is detected, an improved random walk algorithm is used to accurately and efficiently locate the micro-service anomaly. We use the train-ticket micro-service system as a test environment to evaluate the positioning accuracy of this method. The experimental results show that the anomaly detection and accuracy of the method are improved compared with the comparison method.