Hematemesis finishing, performance test methods and steps in detail, the road to advanced testing...

Contents: Guide

foreword

Overall direction: Performance efficiency testing is to use professional load generation equipment from the perspective of user experience to verify whether the system can meet the performance indicators proposed by the user on the basis of the performance model, and whether it meets the performance of the system design in the user documentation focus point.

In the case of normal interaction volume and peak interaction volume of the system, find the performance bottleneck in the system, optimize the software, and finally achieve the purpose of optimizing the system.

The system must not only be able to withstand large concurrent accesses, but also provide users with a better user experience, which means that the system also has high performance requirements. The performance evaluation for the previous stage is also one of the key directions of this evaluation.

Design the performance model by combining the usage habits of the actual system, select typical business points, develop performance scripts and design reasonable scenarios and business ratios according to the actual business requirements of the system, so that the performance evaluation can be based on the actual situation as much as possible. Possible discovery of system bottlenecks, providing reference and basis for system tuning;

test type

Performance testing Test
whether the performance of the system meets the production performance requirements by simulating the business pressure of production operation and the combination of usage scenarios.

Load testing works
by continuously increasing stress on the system under test until performance metrics (such as response time) exceed predetermined metrics or a resource is saturated. This kind of test can find the processing limit of the system and provide data for system tuning.

Stress test
To test the session capacity that the system can handle under a certain saturation state (such as CPU and memory usage), and whether the system will encounter errors

Test indicators

Response time
The time required to complete a transaction.
For example, if it takes 1 second from clicking the login button to returning to the login success page after the login is completed, then it is said that the response time of this operation is 1 second.

In the performance test, the statistics of the response time are completed through the transaction function. The transaction refers to the operation of doing something. The transaction function will record the time difference between starting to do this thing and finishing the thing. Use Transaction Response Time Words to describe, also known as transaction response time.

Throughput
The number of transactions processed per unit of time.
For example, it takes 1 second for a user to log in to the system. If the system supports 10 users to log in at the same time, and the response time is 1 second, then the throughput of the system is 10 per second.

In performance testing tools, throughput is also called TPS (Transaction Per Second, the number of transactions per second), which means the number of transactions that can be completed in a unit of time. TPS is generally calculated by dividing the number of transactions passed by the time.

Server resource utilization
The resource utilization of the system under load.
The lower the resource usage, the better the system.

Resources do not only refer to the hardware that runs the system, but all the software and hardware platforms that support the entire system to run programs.

In performance testing, we need to monitor the occupancy of various resources on the hardware or software of the system under load, such as CPU occupancy, memory usage, query cache hit rate, etc.

Applicable method

The benchmark test method
uses a single virtual user to visit a single typical business point in turn to ensure that there is no other operational impact during the test, so as to obtain the "benchmark" response time, resource utilization, throughput reference value and performance periodic performance trend .

The load testing method
continuously increases the pressure on the system under test until the performance index such as response time exceeds the predetermined index or a certain resource has reached a saturated state. This kind of test can find the processing limit of the system and provide data for system tuning.

The stability test method
allows the system to run continuously for a period of time under the condition of loading a certain business pressure on the system (such as: the utilization rate of resources is 70%-90%), and tests whether the system can be stable under this condition run

test steps

1. Clarify the real needs of users for system performance, and grasp the access indicators that the system is expected to withstand when providing services to the outside world (such as: average user visits, peak user visits, response time required, transaction throughput, etc.).

2. According to the system design documents and communication of user needs, understand the overall system architecture, system business process, system topology, system data flow and other technical information, and conduct basic analysis on it, and preliminarily determine the performance bottleneck point in the system.

3. Create a performance test model. The performance test needs to be based on certain prerequisites. A certain performance is closely related to all aspects of the prerequisites. The performance test model is created by analyzing test requirements and system analysis. Constraints that help limit performance test results sexual conditions.

4. Develop performance test cases according to the test method, develop performance test scenarios and scripts, perform benchmark tests, load tests and stability in sequence, and record relevant performance test indicators and resource utilization.

The following is the most complete software test engineer learning knowledge architecture system diagram in 2023 that I compiled