1. Performance test (performance test)
usually collects all the performance related to the test, and is usually used by different people in different occasions. Test the performance of the software in the system, and measure
the gap between the system and the predefined goals.
2. Load Test (Load Test)
Load test is a kind of performance test, only if the data runs in an overloaded environment, whether the program can bear it. By gradually increasing
the system load, determine the maximum load that the system can bear under the condition that the performance indicators are met. (I don’t know how many users the system can satisfy in the demand, we need to find it)
3. Stress test (Stress Test)
Stress test is a high-load test, that is to say, the system is under a load, and then Continue to pressurize it, forming a double load, until the system crashes, and pay attention to the recovery ability of the system after the crash, pay attention to a process of repressurization, and
see if the system has been completely destroyed.
The external load is called pressure, and the internal pressure is called load. Load focuses on the internal and system itself; stress pays more attention to the external appearance of the system.
The performance test is performed from light to heavy, gradually putting pressure on the system. Usually the performance indicators that users are most concerned about include: response time and throughput. Resource utilization and maximum number of users
can be divided into light load area, heavy load area and load failure area
Light load area: as the number of virtual users increases, the system resource utilization and throughput also increase, while the response time is not particularly Obvious changes
Heavy load area: As the number of virtual users increases, the system resource utilization increases slowly, and the throughput begins to increase slowly. As the number of virtual users increases, the resource utilization remains relatively stable (satisfying the system Resource utilization index), the
throughput is basically stable, and the follow-up is slightly reduced, but the magnitude is not large, and the response time will have a relatively large increase.
Load failure area: In this area, the system resource utilization increases and reaches saturation, the incoming CPU utilization reaches 95% or even 100%, and maintains this state for a long time, while the throughput drops sharply and the response time increases significantly (ie: the occurrence of inflection point)