In the current software testing industry, proficiency in performance testing is already a stepping stone for test engineers to interview. Of course, there are still many testing friends who work more every day. In terms of performance, they may have only done simple concurrent testing. For Writing scripts and setting up the environment are also relatively unfamiliar. Today's article will show you how to do performance testing and how to master performance testing. The article structure is as follows:
How to do performance testing?
Indicators and indicator analysis that testers focus on in performance testing
Advanced Performance Testing
How to do performance testing?
Test tool selection:
It is recommended that you choose RunnerGo, which is simpler than jmeter and more suitable for getting started quickly.
Confirm requirements:
What is the purpose of the test? What is the system under test? Which is the test scenario? What are the performance indicators? It is enough to mainly determine these points, and then write scripts for these points.
Test environment preparation:
RunnerGo does not need to be run on the compressor, you can directly log in to their official website and register to use it directly. Attach its official website address https://www.runnergo.com/
Script design:
We determine some scenarios and interfaces and some performance indicators for performance testing according to requirements. Here RunnerGo supports importing interfaces from various Api debugging tools, and one-click import can be used directly.
After the interface is imported, RunnerGo can directly import the interface, and configure the scene by connecting lines on the background canvas. An interface flow is a scene.
We determine the pressure test mode according to the requirements: RunnerGo has built-in various pressure test modes, and you can configure the desired parameters after one-click selection.
After selecting the stress test mode, a simple performance test script is configured, and we can start testing.
After RunnerGo starts the test, you can view the test report in real time. When the script is running, we can analyze the test report according to the performance indicators. A simple performance test is completed here.
Indicators and indicator analysis that testers need to pay attention to in performance testing:
Indicators that testers need to focus on:
Concurrency, throughput (tps/qps), response time, error rate.
Whether the performance of the pressure-generating machine is normal during the test, and whether the resources of the machine where the stress-tested program resides are normal.
indicator anaysis:
We know that with continuous improvement of concurrency, the system will become more and more unstable, the response time will fluctuate more and more, the response time will become slower and slower, and the throughput will become more and more difficult. The same is true for CPU usage, and the interface error rate will also increase. In this case, throughput is meaningless. Therefore, we need to find the pressure point where the system runs stably during the test, and optimize and improve on this basis. When we issue a test report, we need to ensure the accuracy of various test data, and give more information about the analysis, comparison results, and details of how to obtain the results.