Performance engineering is divided into four stages according to different contents and purposes, namely, the offline single system stress test and analysis stage, the offline full-link stress test and analysis stage, the production read-only business stress test and capacity assessment stage, and the production read and write business. Full-link stress testing and capacity assessment stage. (It can also be understood as the development stage of an enterprise performance testing system)
Offline single system stress test analysis stage
Performance testing for a single system is distributed throughout the software life cycle, including performance testing of a single module during the development phase, performance capacity evaluation of a single system before going online, and system stability monitoring after going online . In summary, this stage evaluates the performance of a single system and performs location analysis on performance problems, and is completed in an offline environment.
The offline single system stress test analysis phase usually focuses on the performance indicators of commonly used or important functions in a single system. Taking the login function interface as an example, the indicators of concern in the offline single system stress test analysis phase include how many users can be logged in by the interface at the same time, how long it takes to log in at the same time, the resources of the server where the system is located in each scenario, and whether users can all succeed. Login etc.
During this process, if the indicator data is found not to meet the needs, problem analysis, positioning and optimization need to be carried out.
There will also be scenarios where multiple functions are tested simultaneously in a single system. In this case, the indicators that need to be paid attention to are similar to the above indicators, but are relatively complex. No matter what the situation is, the core content is to perform stress test analysis on only a single system in an offline environment.
Offline full-link stress testing analysis stage
The offline full-link stress testing and analysis phase conducts stress testing and analysis on system groups or microservice architecture systems of different businesses to obtain the overall performance. It pays more attention to the direct performance impact between systems and services. Because the performance of other systems or services will directly affect the performance of the entire system group or the entire business, only analyzing the performance of a single system or single service can no longer meet the requirements of performance engineering. In short, the focus of this stage is the performance of the entire link. Offline full-link stress test score