I am interested in algorithm that calculates p(90), p(95). These values are displayed after tests run.
Could you tell me the algorithm name, wiki page or where I can take a look piece of code if it in open source?
The k6 code that calculates percentiles is found here: https://github.com/loadimpact/k6/blob/master/stats/sink.go#L113-L131
And although that’s currently not the same code that’s used to calculate the percentiles in the Load Impact cloud execution (since in it, multiple machines could execute the same script), what we do there is conceptually similar - after all, percentiles don’t have many varied interpretations.
1 Like