Jmeter aggregation Report-90% line parameter description

In fact, it is very simple to describe the meaning of this parameter. You may have known it for a long time, but I have always misunderstood this parameter and thought it was "truth ", I was asked this question in an interview, which caused a new understanding of this parameter.

 

Let's talk about my incorrect understanding:

 

I keep90% lineIt is understood:90%The average response time of the user.

Assume that10Quantity:

1 , 2 , 3 , 4 , 5 , 6 , 7 , 8 , 9 , 23

I will remove it from this group.10%That is, remove23Because if23When calculating the average value, there will be a large error. Most (90%) The user's response awareness is so bad.

If the average response time of all users is calculated as follows:

(1 + 2 + 3 + 4 + 5 + 6 + 7 + 8 + 9 + 23)/10 = 6.8In seconds, the average user response time is6.8Seconds.

Remove% 10(23) Then calculate the average response time:

(1 + 2 + 3 + 4 + 5 + 6 + 7 + 8 + 9)/9 = 5In seconds90%The average response time of a user is5Seconds.

 

Then, let's assume the number of the following two groups:

1 , 2 , 3 , 4 , 5 , 6 , 7 , 8 , 9 , 10

5 , 5 , 5 , 5 , 5 , 6 , 6 , 6 , 6 , 6

I beg90%The average response time of, which number should I go? It seems that there is no "alternative" number that is too prominent.

 

Why does this problem occur? Because I see a lot of information (not on the official website)90% line"Is interpreted:90% user response time

 

 

90% the correct meaning of the line parameter: 

 

Although there is a certain truth in my understanding above, it is obviously wrong. Let's see.JmeterWhat does the official website say?

 

90% line-90% of the samples took no more than this time. The remaining samples at least as long as this.

 

"90%The sample does not exceed this time, and the remaining samples only need this."(TakeGoogleTranslated)

I didn't quite understand what it meant. So, click to explain in detail.

90% line (90 th percentile) is the value below which 90% of the samples fall. the remaining samples too at least as long as the value. this is a standard statistical measure. see, for example:Percentile entry at Wikipedia.

The English is too bad and I still don't understand what it means, but I am prompted to use Wikipedia to check what is a percentile.

 

Percentile:

Statistical Terminology. If a group of data is sorted from big to small and the corresponding cumulative percentile is calculated, the value of the data corresponding to a one hundred-bit value is called the percentile of the one hundred-bit value. Can be expressed as: A group of NObservations are arranged by numerical values, for exampleP %Location ValuePPercentile.

The median value is 50th.Percentile.

25th The percentile is also called the first quartile ( First quartile ) , Use Q1 Number 50 The percentile is also called the second quartile ( Second quartile ), Use Q2 Number 75 The percentile is also called the third quartile ( Third quartile ) , Use Q3 . If P The percentile is a decimal number, which can be a complete integer.

Quantile is a measure of the location of a data, but it is not necessarily a central location. The percentile provides information about how each data item distributes between the minimum and maximum values. For data that does not have a large number of duplicatesThe percentile divides it into two parts. AboutP %The value of the data itemPThe percentile is small.(100-P)% Value ratio of data itemsPThe percentile is large. ForPThe percentile, which is strictly defined as follows.

PThe percentile is a value that causes at leastP% Of data items are less than or equal to this value, and at least(100-P)The value of % is greater than or equal to this value.

The admission exam scores of higher education institutions are often reported in percentages. For example,Assume that the original score of a candidate in the language section of the entrance exam is 54.Points. Compared with other students taking the same test, it is not easy to know how their scores are. However, if the original score54Which exactly corresponds to70Percentile, we can know about70% Of students have lower scores than him, while about30% Of students scored higher than him. 

 

In fact, this is the final example of the examinee's score, which makes me suddenly understand.

 

Suppose:

There are 10:

1,2,3,4,5,6,7,8,9,10Sort them in ascending order.

Please calculate its 90%Percentile,That is,9The number is exactly9So his 90% lineYes9.

Number of other groups:

2, 2 . 1, 2 . 5, 3 , 3 . 4, 3 . 4, 4 , 4 , 4 , 4 , 5 , 5 , 5 , 5.9 , 5.91 , 6.8 , 8 , 12 , 24 , 24.1 Sort them in ascending order.

Please calculate its 90%Percentile, number18 is 12He's90% lineYes12. 

 

90% line

Sort the number of a group from small to large and find its90%Number (for example12), Then the array contains90%Is less than or equal12.

The response time used for performance testing also makes sense, that is90%The user response time will not exceed12Seconds.