Analyzing a Jetty request file via command line

Issue

Sometimes we need to analyze a Jetty request log quickly to see more about what is happening in a time series. Through the power of awk, we can see some data relatively quickly from the command line. This gives a per minute time series of average time in ms, total time taken across requests, max time and total number of requests during that minute.

Environment

Jetty, Linux

Resolution

This is an example querying about '/select' traffic through a component. 

The fields output are the minute, avgTime, avgBytes, maxTime, totalCount:

cmd> grep '/select' jetty-2020_10_01.request.log | awk '{print $4, $NF, $(NF-1)}' | cut -c 14-17,21- | awk '{arr[$1]+= $2; b[$1]+=$3; count[$1]++; if ($2 > max[$1]) max[$1] = $2} END {for (key in arr) printf("%s\t%6.2f\t\t%6.0f\t\t%i\t\t%i\n", key, arr[key] / count[key], b[key] / count[key], max[key], count[key])}' | sort

00:00 55.78 121743 129 106

00:01 58.00 114417 181 101

00:02 60.98 141868 197 128

00:03 52.65 126631 104 96

00:04 57.03 129881 149 93

00:05 54.74 141087 107 122

00:06 55.64 139846 212 121

00:07 54.48 138532 129 120

00:08 53.48 119547 150 83