When will the performance test be carried out? What is the purpose of performance testing?

Date：2022-03-25 15:14:58Views：807

The basic principle of performance testing is: the earlier the problem is found, the easier it is to locate the problem and repair the problem. Therefore, performance testing can run through all stages of the software development life cycle. For example, unit testing can test the functional accuracy under multithreading concurrency, each API also needs to be tested and evaluated, the data size and cache usage of the modules used need to be taken into account in the integration test, and the relevant global performance indicators need to be measured from the perspective of user load in the system test. So, when will performance testing be carried out? What is the purpose of performance testing?

1. When will the performance test be carried out?

Performance detection is based on the fact that the functions of the system have been completed or tend to be completed. If the functions are not complete, it is of little significance. Because the later function improvement will affect the performance of the system, entering the performance test too early will lead to inaccurate test results and waste test resources. Therefore, performance testing is first based on functional testing. Performance testing can only be carried out after knowing its functional requirements.

2. How to test performance?

For a system under test, we need to analyze it in three parts:

Entrance: how to send the request, how much pressure the pressure party should exert, and how to exert pressure;

System under test: how the system responds to a single request, system business process, system network element nodes, data flow, etc. Whether the overall performance requirements exist, which indicators need to be checked and how to monitor;

Export: what data is received and how to obtain and compare it;

Isn't there much difference between sensory and functional tests? Yes, it is to analyze the function flow of a single user and the data flow structure diagram of the system (including the data flow in the background), and then consider a large number of user operations.

The general performance test steps of the system are as follows:

1. Confirm the test objectives;

2. Analyze the business requirements of the test system;

3. Analyze the system structure of the tested system;

4. Analyze the performance test points of the tested system;

5. Design test scheme, test scheme and test case;

6. Select test tools;

7. Testing and development;

8. Test execution;

9. Analysis of test results;

10. Test tuning, test verification and test analysis;

11. Output test report;

Performance testing has different purposes in different situations. Knowing the objectives of each performance test can make us more targeted when carrying out performance testing.

Purpose of performance test

1. Obtain the performance indicators of the system under normal request traffic:

This is the most common purpose of performance testing, which is usually required for routine performance testing. Test the performance of the system under normal pressure (no pressure) flow, and establish a measurable reference standard to provide comparative reference for other test scenarios or tuning processes.

Common performance test indicators include:

Service response related: average response time, response time distribution, failure rate, etc

Resource usage related: CPU usage, memory, disk IO, network IO, etc

2. Obtain the ultimate capacity that the system can carry and the performance bottleneck of the system:

Because the system capacity evaluation is difficult to automate, the frequency of such requirements is usually lower and may be evaluated periodically. Through the "inflection point method" for performance test, gradually increase the pressure of the system until the performance index is unacceptable or there is an obvious inflection point, so as to find the bottleneck of the system. It should be noted that the limit concurrency or the limit number of requests processed per second during pressure test must be based on the performance standard of safe operation and maintenance.

3. Test system stability:

System stability pressure measurement is to know whether the system can operate stably under various conditions.

The conditions causing system instability include:

Limit requests per second

Maximum number of concurrent or connections

Sudden traffic peak

Long time pressure flow

Hotspot data request

Poor network environment

Slow return of downstream module, timeout or false death of fault, etc

4. Parameter adjustment of auxiliary system for performance test:

We need to cooperate with R & D to test the comparison of system performance under different system parameters or deployment schemes.

5. System performance regression:

When upgrading a version, you usually need to compare the performance of the new and old versions.

After repeated performance testing, users can have a better product experience in any environment. The ultimate goal of performance is to ensure that the product meets the performance and enable users to have a better experience.