Performance testing purposes : to find performance bottlenecks
Performance Testing Category :
- Test Load : stepwise by pressure, to achieve a predetermined performance target threshold value is set to be smaller than the threshold value is equal to a certain value, such as 80% or less cpu utilization.
- Stress Test : by stepwise pressure, making some of the resources of the system is saturated, even a failed state, simple and crude explanation is what conditions the system pressure can collapse.
- Concurrent test : at the same time, multiple virtual users simultaneously access the same module, agreed to function, the usual method is to set up a collection point test
- Capacity Test: usually refers to the database level, the goal is to obtain optimum capacity database capabilities, also known as the capacity estimated specific methods: In certain of concurrent users, different amounts of basic data processing capability of the observation database, That acquisition performance indicators database
- Reliability Test: also known stability tests or long fatigue testing is the system running the system is stable at high pressures, such as above 80% cpu utilization, 7 * 24-hour operation, the system is stable....
- Abnormal test system reaction is also known as failed test means the test system architecture, and load balancing architecture, to test down, hang nodes, and other situations:
Performance testing workflow:
- Needs analysis ---> --- to develop performance indicators> Script Development ---> scene settings ---> Monitor deployment ---> test execution ---> Performance Analysis ---> Performance Tuning --à test report
Common system applications layered architecture:
- Display layer (view) web, android, ios, H5
- Logic control layer (controller) API
- Data storage layer mysql, mongodb, redis
Transaction:
- From one or more requests initiated by the client (consisting of a request to complete these operations) to the client receives a response back from the server
Tsp(transaction per Second):
- Number of transactions per second, the system can handle
Request response:
- The client initiates a request until the client receives the response back from the server, the process time-consuming.
Transaction Response Time:
- A transaction may be composed of one or more requests, the transaction response time is mainly directed to the user's point of view, such transfer.
Concurrency is defined:
- No concurrent in the strict sense, there has always concurrent, whether the gap is 1 millisecond or microsecond, there is always a time lag, so the concurrent tells of a time frame, such as one second
- Concurrent Example 1: Multi-user operating on the same system, such as dual eleven, they have carried out the same kind of goods for spike
- Concurrent Example 2: multi-user operating on different systems, such as when two-eleven, we carried out for different commodities spike, or we have carried out various other operations, such as browsing merchandise
Number of concurrent users:
- The number of users within the same unit of time, the system initiates the request
Throughput:
- A sum of the performance testing process, the amount of data transmitted over the network
Throughput:
- Per unit time, the amount of data transmitted over the network
- Throughput = throughput / transmission time
Hits:
- Per number of requests submitted by the user to the server, this ratio is an indicator web application-specific, imagine how many clicks on the page carried a total of users per second, but note that a mouse click operation after the client may have sent multiple requests like server.
Resource Utilization:
- Use of different system resources, such as cpu, memory, io
Performance Testing Requirements Analysis:
- Objective: To determine test indicators; clear test scenarios;
- The new system: compared with the industry; business expectations;
- Old system: Contrast the previous user behavior and user volume