jmeter-statistics

A CLI written in node for analysing jmeter jtl files. You can pipe a csv formatted jtl file into it, and it will output a .csv with aggregated data.

Installation

> npm install -g jmeter-statistics

Usage

The following commmand will analyse the results.jtl file and produce results into a comma separated file (csv).

> cat results.jtl | jmeter-statistics > statistics.csv

The csv will contain output similar to the following

> cat statistics.csv

label,requestCount,meanResponseTimeMillis,maxTime,minTime,errorPercentage,apdex,satisfied,tolerating,frustrated,ninetiethPercentile
label1,2,1838,2000,899,0.25,0,1,1,1900
label2,5,9566,11982,8746,35.57,10.29,0,1,9520,46,0,11982

Fields

• label: the label as defined in the jmeter test
• requestCount: the total number of requests made
• meanResponseTimeMillis: the mean response time in milliseconds
• maxTime: the maximum time in milliseconds
• minTime: the maximum time in milliseconds
• errorPercentage: the percentage of requests that resulted in a non-200 response code (as a value between 0 and 1)
• apdex: the apdex score based upon a target time (T) of 500ms
• satisfied: the number of requests defined as satisfied by the apdex target time (response time <= 500ms)
• tolerating: the number of requests defined as tolerating by the apdex target time (response time <= 2s)
• frustrated: the number of requests defined as frustrated by the apdex target time (response time > 2s)
• ninetiethPercentile: the value below which 90% of the samples fall (See explanation)
• p50Percentile: The median request response time
• p99Percentile: The p99 response time (ie: only 1% were slower than this figure)