I want to track API response time. I can use methods such as medium, medium and others for monitoring. But I am having the following problems with these methods:
Medium Problem
if one of the requests takes a very long time. For example, in a given preset tool, it will become high due to a value of 1000.
S1 = [1, 1, 1, 2, 2, 2, 2, 2, 2, 2, 2, 1000]
Problem with the med
This will be the correct value only up to 50%. For example, in a given set S2 = [2,2,2,2,2,50,50,50,50]. the median gives a value of 2, but most users experience a slow response.
Problem with 5-95 span ( http://steveakers.com/2013/08/01/span-vs-median-for-response-time-monitors/ )
In the above article, the author suggested using the value uppser95-uppser5. But this will not generate a warning if the response time is similar: s3 = [50,50,50,50,50]. In this case, all API responses are slow. But the range of 5-95 is zero.
I am thinking of using either of these two values. upper95 or (upper95 + upper5) / 2.
Which one will be better and why? Is there a better way to calculate QOS?