Vegeta is a versatile load testing tool which can be used both as a command line utility and a library.

The key vegeta features I enjoy are:

• A simple easy to use, command line interface
• A config file format which trivially allows setting custom headers, which is key to working with pre-production services, and allows testing multiple endpoints simultaneously
• The ability to run multiple tests in parallel across different hosts, increasing the total test volume possible
• It generates a constant request rate, a feature found in few other tools, but also a key to finding true operational limitations.
• Since it’s written in Go, it’s compiled to a single binary, making it trivial to drop into machines if you’re doing scale out or remote node testing.
• It also has an excellent library interface, allowing it to be a component of custom tooling.

brew update && brew install vegeta

$ vegeta attack -duration=1s -rate=200 -targets=data/urls-with-methods.txt -header="Authorization: Bearer $token" | vegeta report --type=text

Invokes the performance test runner and allows you to specify the file whose content will be set as the body of every request

Specifies the amount of time you want the tests to run. The CLI allows you to pass in a variable (example in above command). You can pass in 0 as an integer for infinite runtime of your test scenario.

Specifies the request rate per test run. If you don't specify the rate through the CLI, then 1s becomes the default. To give you an example on usage for our specific load scenario above, we're executing 200 concuurent requests per second.

This is where you can specify which endpoint you want to test. Typically you will specify the http method & the endpoint being tested:

GET https://endpoint

Every load test that is executed will write the results out to stdout in text, which can be specified using the --type CLI flag. The results from the run will display in your terminal:

Requests      [total, rate, throughput]         200, 201.09, 6.49
Duration      [total, attack, wait]             30.666s, 994.567ms, 29.671s
Latencies     [min, mean, 50, 90, 95, 99, max]  2.855s, 4.391s, 4.286s, 4.969s, 5.113s, 6.15s, 30.006s
Bytes In      [total, mean]                     8379933, 41899.67
Bytes Out     [total, mean]                     0, 0.00
Success       [ratio]                           99.50%
Status Codes  [code:count]                      0:1  200:199