The Spring Festival has just ended, and careful friends found that jmeter 4.0 was also quietly launched. Combined with everyone's usual usage, under the analysis of the features added to the new version, the download address of jmeter 4.0: http://jmeter.apache.org/download_jmeter. cgi
1. First download and open jmeter 4.0, say a visual change that can be felt, as shown in Figure 1 below, the black interface, many friends think that the technical black things look higher and bigger, although this concept is a bit superficial, but it seems It makes sense, after all, there are still many friends who are keen on the dark interface of linux.
figure 1
2. Jmeter 4.0 supports java 9 version. With the popularization of java 9, the incompatibility of the old jmeter version is still a rather painful problem. After all, most people do not want to install two java versions or frequently switch environment variables;
3. Regular expression expression optimization, I think it is optimization, although 4.0 has been renamed Boundary Extractor, but still retains the original Regular Expression Extractor, in fact, after careful comparison, it is found that, as shown in Figure 2, Figure 3,
figure 2
image 3
In addition to some changes in the regular part, others are basically the same. The new version allows users to fill in more popular, which makes it easier for novices to understand and get started. Another reason given by the official jmeter is with better performances, that is, better performance .
4. Added json assertion. For the old version, there are also external plug-ins. 4.0 integrates this. Let me give you a chestnut. The returned data is shown in Figure 4. I judge the username according to the userid, and the specific verification format is shown in Figure 5. , you can figure out the format according to the case. Novices may be prone to mistakes. Remember to check the green part. Some data information in red is hidden. I will copy the json path format. $..DataList[?(@.UserId==' xxxxxxx')].UserName
Figure 4
5. Precise Throughput Timer , first let you know how to control it, as shown in Figure 5, the first line to be filled in Target throughput (in samples per "throughput period"), you may think at first glance that this line is to control tps, in fact, you need to combine In the second line, as shown in the figure, only one request is allowed within 10s, and the real tps is controlled at 0.1.
Figure 5
Jmeter also launched Constant Timer and constant throughput timer before. I think this shows that Jmeter officials have spent a lot of energy on comprehensive scene control. The core scene of performance test is also comprehensive scene. For the control of comprehensive scene, in version 2 At times, it is still not comparable to LR, including that LR can freely adjust virtual users during the execution process, which is currently unreachable by JMETER. Through some instructions, we can know that jmeter still controls tps through the thinking time between threads. This method is slightly rough, but with With jmeter's investment in this area, I believe that the gap will be smaller and smaller, or there will be more alternatives.
Jmeter 4.0 abandons the workbench method for jms message viewing and recording. Recording is rarely used for me. Interested friends can try new recording plug-ins. In addition, 4.0 fixes the bugs left over from version 3.3. , 4.0 generally gives me a good feeling, and I hope that you can try more. Of course, there will be new bugs in the launch of the new version. For companies that require high performance testing accuracy and stability, you may wish to slow down, etc. After the version is stable, it will be promoted for use.