Cloud resource topology anomaly detection method and system

Abstract

The application provides a cloud resource topology anomaly detection method and system, the method comprises: acquiring node data and edge data of a cloud resource configuration management database, constructing a node feature matrix and an edge index, the node feature matrix is composed of numerical features and device type codes of nodes; inputting the node feature matrix and the edge index into a graph autoencoder to generate node embedding vectors and calculate feature reconstruction errors of each node, and determining abnormal nodes according to the feature reconstruction errors of each node and a preset anomaly threshold. The application can improve anomaly detection accuracy by using the device association relationship of the cloud resource topology, adapt to the differences in normal operation characteristics of different types of devices, and reduce the false positive rate and the false negative rate of anomaly detection.

Images