Cloud Platform Component Fault Tracing System Based on Operation and Maintenance Knowledge Map
Yongyan Yang, Lihong Yang · 2022
Conventional cloud platform component fault traceability system uses graph convolutional neural network method to extract component fault event node time and location information. The conventional method has low accuracy and a complex extraction process, which leads to a long time-consuming system cloud platform component fault traceability. This paper introduces operation and maintenance knowledge mapping to address this problem and proposes a new fault-tracing system. The system hardware adopts B/S architecture, which can guarantee the reliability and efficiency of system operation. Firstly, this paper constructs a component fault operation and maintenance knowledge graph of the cloud platform and extracts the time and location information of component fault event nodes. Secondly, this paper designs the component fault traceability analysis model obtains the dynamic changes of equivalent fault rank, and completes the component fault traceability by combining the component fault event information. The system test results show that the cloud platform component query and fault traceability analysis takes less time with the traceability system designed in this paper, which can quickly achieve the cloud platform component fault traceability goal.