wrktoolbox

A tool to run wrk and wrk2 benchmarks, store their output, and generate reports.

Features

• Support for YAML and JSON configuration files, to define benchmark suites
• Parses the output of wrk and wrk2 HTTP benchmarking tools
• Possibility to define performance goals, which are evaluated and stored with results
• Strategy to store benchmarks results and whole suite configuration
• Support for plugins, loaded dynamically to define new types of stores, performance goals, and reports writers
• Strategy to produce reports of results, for example to XLSX, with wrktoolbox-xlsx - see Wiki
• Docker images for Ubuntu and Alpine
• CLI

Quick example

1. install

pip install wrk-load-tools

2. verify that the cli works (optional)

wrktoolbox --help

3. prepare a YAML, or JSON file with configuration. A basic example, with a single type of benchmark, looks like this:

# the array of benchmarks contains the configuration of benchmarks to run
benchmarks:
  - test_id: this_is_an_example_test_id #preferred test identifier to be added in reports
    url: https://this-is-an-example.it/api/alive
    threads: 10  # threads count
    concurrency: 100  # concurrent users
    duration: 30  # test duration in seconds

# the type of stores to use, to collect benchmark results
# it is possible to define custom stores, using plugins;
# for example to store results in a database, or send them to an API
stores:
  - json

4. run a suite of benchmarks using a settings file

wrktoolbox run --settings basic.yaml

Refer to examples folder for an example of full configuration file, defining plugins for authentication and custom store.