Request response state determination method and device, equipment and storage medium

Abstract

The embodiment of the invention discloses a request response state determination method and device, equipment and a storage medium. The method comprises the following steps: acquiring request information; performing vectorization processing on the request information, and determining a target feature vector sequence; inputting the target feature vector sequence into a predetermined target state prediction model, and determining the response state of the request information according to the output of the target state prediction model; the problem of difficulty in fault analysis and positioning is solved, the target state prediction model is trained in advance, the request information is vectorized to obtain the target feature vector sequence, the target feature vector sequence is input into the target state prediction model for response state prediction, and the response state of the request information is predicted in advance. The situation that the fault cannot be accurately positioned due to the fact that the number of transaction data is large after the fault occurs is avoided, the response state of each piece of request information is automatically predicted, the workload of workers is reduced, and the working efficiency and the system reliability are improved.

Description

technical field [0001] The embodiments of the present invention relate to the field of computer technology, and in particular, to a method, device, device, and storage medium for determining a request-response state. Background technique [0002] More and more core systems use distributed architecture to move down to open platforms. Distributed architecture has many advantages, but there are also many difficulties that cannot be ignored. During the operation of the distributed system, due to the complex call relationship in the distributed system, when a node fails, the transaction in the system fails. If a failure occurs during performance testing, it will be wrong if the failure is not found in time Information will quickly fill the log, affecting the function of the entire system, and even causing the system to crash. [0003] At present, fault analysis and failed transaction analysis of distributed systems usually locate faults through full-link monitoring, Dynatrace or...

AI Technical Summary

Problems solved by technology

Dynatrace can monitor performance, but when the performance of microservices is poor, it is still necessary to manually locate the node where the problem occurs and analyze the logs of each node, which increases the workload of the staff and reduces work efficiency and system reliability.

Images