Average Question
If one of the requests takes a long time. For example, in a given set, due to the value 1000, the average value will become higher.
S1 = [1,1,1,2 ,2,2,2,2,2,2,1,1000]
Median problem
Its correct value is only as high as 50%. For example, in the given Set S2 = [2,2,2,2,2,50,50,50,50]. The median value gives the value 2, but most users face a slow response.
5-95 Span (http://steveakers.com/2013/08/01/span-vs-median-for-response-time-monitors/) issues
In the above article the author recommends using the value uppser95- uppser5.
But if the response time is as follows, no alert will be generated:
s3 = [50,50,50,50,50]. In this case, all API responses are slow. But span 5 -95 is zero.
I am considering using either of these two values.
upper95 or (upper95 upper5)/2.
Which one is better and why ? Is there a better way to calculate QOS?
> Average (average) response
>Median response
> 5-95 span response
Please note that the measured values of #3 and #1 and #2 are different!
>Average and median can measure the actual response time. This will cause certain types of problems.
> 5-95 span tells you the degree of change in response time. That is, whether your response time is consistent This will cause another type of problem.
You may need to track both: absolute response time and variance. The best method for the former (mean and median, whether to cut out outliers) may depend on For you to get the results of the service.
I want to monitor the response time of the API. I can use the average, median and other methods to monitor. But I am facing the following problems:
Average problem
If one of the requests takes a long time. For example, in a given set, due to the value of 1000, the average will become higher.
S1 = [1,1,1,2,2,2,2,2,2,2,1,1000]
The median question
Its correct value Only up to 50%. For example, in a given set S2 = [2,2,2,2,2,50,50,50,50]. The median value gives the value 2, but most users face slow responses .
5-95 span (http://steveakers.com/2013/08/01/span-vs-median-for-response-time-monitors/) issues
In the above article the author recommends using the value uppser95-uppser5.
But if the response time is as follows, no alert will be generated:
s3 = [50,50,50,50,50]. In this case , All API responses are very slow. But the span 5-95 is zero.
I am considering using either of these two values.
upper95 or (upper95 upper5)/2.
p>
Which one is better and why? Is there a better way to calculate QOS?
You listed three measurements:
>Average (average) response
>Median Response
> 5-95 Span Response
Please note that the measured values of #3 and #1 and #2 are different!
>Average and median can measure the actual response time. This will cause certain types of problems.
> 5-95 span tells you the degree of change in response time. That is, whether your response time is consistent This will cause another type of problem.
You may need to track both: absolute response time and variance. The best method for the former (mean and median, whether to cut out outliers) may depend on The result of your service.