Micro-application system anomaly detection method and device based on graph neural network

The invention relates to the technical field of network communication security, and discloses a micro-application system anomaly detection method and device based on a graph neural network, and the method comprises the steps: obtaining a call chain and log data of a micro-application system, the call chain comprising a span event, and the log data comprising a log event; performing feature learning on the span event and the log event by adopting a large language model to obtain event feature vectors corresponding to the span event and the log event; constructing a heterogeneous event graph based on the event feature vector, the call chain and the log data, wherein the heterogeneous event graph is used for representing the relationship between the span event and the log event; training a graph neural network by using the constructed heterogeneous event graph; and performing anomaly detection on the micro-application system by using the trained graph neural network. The problem that micro-application anomaly de