k6 Test Results Visualization

• a simple tool written in Python for visualization of k6 output
  • when passing --out json=<file_name> to the k6 program all test events are serialized and saved to the given file
  • this file can be later processed and its data visualized
• in the current state it allows for plotting given metric (it's average and maximum value in each minute/second)
  • by default it plots the http_req_waiting metric values

Motivation

• seeing how certain metrics' values change during a load test is quite important as it can provide useful information
  • take average response time changing as the number of VUs increase as an example
• running the k6 tool locally does not offer any visualization
  • all tests run in the k6 cloud offer various charts to be analyzed, but sometimes, when running small load tests, we do not want to use paid features of the cloud
• to offer free visualization options for local k6 load tests this project was created

Dependencies

• Python 3 interpreter
• matplotlib (https://pypi.org/project/matplotlib/)

Usage

• run a k6 test and save its output to a file
  • k6 run --out json=k6_test_results.json test.js
• python3 k6_plotter.py k6_test_results.json
  • plots the http_req_waiting metric data by default
• python3 k6_plotter.py <metric>
  • plots the data of the metric identified by it's name (<metric>)

TODO

• better argparsing
  • pass file to which the chart should be saved in SVG format
  • toggle whether date should be rounded to seconds/minutes,... (minutes by default right now)
• optimizations
• finish displaying of a histogram
  • arguments to toggle this feature are missing right now