Method and device for positioning service timeout and storage medium

Abstract

The invention discloses a method and device for positioning service timeout and a storage medium, and the method comprises the steps: recording the time when a request reaches a first service, and forwarding the request to a proxy server; recording the time when the request arrives at the first proxy server, and forwarding the request to a second service; if the second service is the target service, recording the time when the request reaches the second service, and returning a request response; recording the time when the request response reaches the first proxy server, and returning the request response to the first service; recording the time when the request response reaches the first service, and returning the request response to the client; recording the time when the request response arrives at the client; time comparison is carried out, and a service timeout position is positioned. The method and the device are used for realizing quick positioning of the timeout problem occurrence position.

Description

technical field [0001] The present invention relates to the technical field of computers, in particular to a method, a device and a storage medium for positioning service timeout. Background technique [0002] The back-end service on the server side usually develops business logic based on a framework (such as spring). Before entering the business logic of a service, the framework receives the request and encapsulates the parameters, and then calls the business code. If the print log records the request entry time, when the request volume is large enough to a certain extent, the printed time will have errors. Because the service itself will be stuck due to the high concurrency performance bottleneck, there will be a delay in the recording time of the service itself, and the inaccurate data cannot be used as the basis for problem location, so it is impossible to judge the time when the request actually enters the back-end service. [0003] For this kind of problem, the most ...

AI Technical Summary

Problems solved by technology

For example, when client A sends a request, when the request reaches service B, the packet capture tool will record the time when the request arrives at service B, and service B continues to request service C. At this time, the packet capture tool will continue to record the time when service C receives the request, and After service C returns the request, the time when service B receives the request and returns is based on the time when service B receives the request from client A, the time when service C receives the request from service B, and the time when service B receives the request from service C. The following three situations are determined by the comparison between the two parties: 1. Due to the request backlog of service B, service C is not requested in time, resulting in a timeout; 2. Service B sends a request, and the request arrives at C, but service C has no time to process it Timeout caused by the request; 3. Service B requested C, and C returned the request within the specified time, but after B received the request, the subsequent processing result data timed out, resulting in failure to return the result to the client in time, resulting in a timeout

[0004] In the aforementioned network card packet capture method, since the packet capture tool will record all the data passing through the network card, when the request volume is extremely large, the data packets captured by the packet capture tool will be very large, and at this time it is necessary to analyze the data packets. By comparing each The time when a request is sent and the time when it is returned is used to locate the problem, resulting in very low efficiency and extremely inconvenient operation.

For example, a Chinese patent with publication number CN111143179A disclosed on May 12, 2020 a method for locating performance bottlenecks, which uses a packet capture tool to perform node capture and statistical analysis at each moment to locate performance bottleneck nodes. Performance bottleneck positioning, but its packet capture method greatly affects work efficiency

Images