Jmeter's aggregate report is a very nice lintener, which is used in interface testing and performance testing.
So where is the excellence? Above
You may only focus on this part of your daily work:
But do you really understand the indicators here? After reading several blogs of well-known masters, they all focused on the content of aggregated reports; of course, during the interview process, I often asked interviewers who wrote "Proficient in Jmeter" on their resumes about aggregated reports. Questions, regrets, very few can answer; or can answer the focus. There may be a certain misunderstanding in one of the indicators in this area.
Although in some blogs and official account articles, some masters use a separate space to explain it, it seems that "Houlang" doesn't care that much. Or maybe the indicators on aggregated reports are really misleading.
So, I'll also focus on the aggregated reporting aspect.
As always, post the official documents first (some people often complain that they can’t understand English, and it’s a flaw, it’s always a flaw, and it’s time to make up for it. For the urgent need, Google Translate, Youdao Translator, etc. are good strategies!):
Each indicator is clearly explained in the help document. Maybe some translation reasons or self-perception reasons have led to some misunderstandings. Let's interpret them one by one.
Label: The popular translation is the label. (This label is usually less controversial)
[The label of the sample. If "Include group name in label?" then add the name of the thread group as a prefix. This allows identical tags to be organized separately from different thread groups, if desired. 】(Baidu translator)
#Samples: [Number of samples with the same label], the number of requests.
Average: Average response time.
1,2,3,4,5,6,7,8---The average response time of this set of data is 45/8=5.625.
Median: Median.
1, 2, 3, 4, 5, 6, 7, 8, 9---the median of this set of data is 5.
90%Line: [90% of the samples do not exceed this time. The remaining samples are at least as long as this one. (90th percentile)] What is the translation?
Some people usually understand the metric "90% Line" as the average response time of 90%.
At this time, a relatively easy-to-understand concept can be derived: as we all know, China is a country with a large population. In order to measure the economic development, relevant statistical departments will conduct statistics every year. , so how is the number of people in this piece calculated?
Of course, it is unrealistic to count one by one (now), then at this time, the 90% Line indicator is effective. Similarly, let's bring it in first. The concept is "the average response time of 90% Line". According to this concept It is completely unscientific in terms of it; the extremely rich cannot be excluded, and the extremely poor cannot be excluded, is it not?
0,1,2,3,4,5,6,7,8,9; 90% of the numbers are 0,1,2,3,4,5,6,7,8,90%Line is (0+ 1+2+3+4+5+6+7+8)/9, this understanding is obviously incorrect.
Then, substitute the concept in translation [90% of the samples do not exceed this time], assuming 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, the value of 90% Line is 8. The following means that no numbers above 8 are "poor", and the reliability is relatively high compared to the average response time.
95%Line: Same as above.
99%Line: Same as above.
Min: Minimum value.
Max: maximum value
Error%: Error percentage.
Throughput: Throughput.
Received KB/sec: Received KB/SEC - Throughput measured in kilobytes received per second.
Sent KB/sec: Sent KB/SEC - throughput sent in kilobytes per second.