Method and device for adaptively locating maximum processing capacity of http service

Abstract

The invention relates to a method and a device for adaptively locating the maximum processing capacity of an HTTP service. The method includes obtaining http requests from nginx logs; copying the environment from the production environment to the docker container; the http request interaction tool obtains data from redis and sends the data to the docker container; and determines whether to continue the test according to the monitoring information of the docker. The device includes an initial data acquisition unit, an initial environment layout unit, an http request interaction tool and a test termination judging unit. The present invention restores the online http request in the docker and deploys the corresponding online application by analyzing the log of the front-end proxy server, completes the initialization of the stress test environment, and then performs the test, obtains the container monitoring information by monitoring the docker, and obtains the maximum processing in the test Ability to automatically expand http request data before.

Description

technical field [0001] The invention relates to the field of performance testing, in particular to a method and a device for adaptively locating the maximum processing capacity of an HTTP service. Background technique [0002] The traditional measurement of the maximum processing capacity of an http service requires a large number of operation and maintenance personnel to deploy the test environment of the application system involved in the http service, and there will be differences in network topology, hardware device configuration, and quantity between the test environment and the production environment. There is an error between the final test result and the actual result. Professional and experienced performance test engineers are needed to write test scripts and perform performance tests. The test cycle is long, the cost is high, and the test results cannot be issued in time. [0003] Existing technologies, such as CPC, a performance testing system based on online req...

AI Technical Summary

Problems solved by technology

This leads to the need to invade the online system when using gor, which may cause problems in the online environment

And because gor's traffic is recorded in real time, it is not easy to intercept traffic during peak hours

When using the CPC system, users still need to manually deploy the stress testing environment, and there is still a relatively large O&M burden

And the CPC system can only execute the request according to the predetermined tps, it needs to obtain the maximum processing of the system to be tested, and needs to adjust the test results multiple times

[0004] Therefore, people in the prior art have the following problems: 1) a large amount of manpower is required to participate in the work of deploying the test environment in traditional performance testing; 2) in the traditional way of intercepting front-end network traffic, the impact on the production environment is reduced , reducing the burden of system maintenance; 3) Testers write test scripts for http requests that need to be tested, which takes up a lot of time for R&D and testers; 4) The performance test of the test environment cannot reflect the processing capacity of the production environment system in a timely manner

Images