Non-intrusive network fault diagnosis and prediction method in cloud native environment

The invention particularly relates to a non-intrusive network fault diagnosis and prediction method in a cloud native environment. According to the non-intrusive network fault diagnosis and prediction method in the cloud native environment, an eBPF-based non-intrusive cloud native application security monitoring technology is adopted, an application monitoring-oriented kernel instrumentation and event triggering mechanism is combined, and network data are collected by setting a network performance monitoring data collector and are preprocessed; and a cloud native environment network fault diagnosis and prediction model based on the Bayesian network is established, monitoring data are analyzed, and diagnosis and prediction of network faults are realized. According to the non-intrusive network fault diagnosis and prediction method in the cloud native environment, network faults possibly occurring in the cloud native environment can be predicted in real time, the occurring network faults can be diagnosed and ana