User Guide

Welcome to the Stressify.jl User Guide. This document will walk you through the basic steps of using Stressify for performance testing.

Stressify Performance Testing was created to facilitate the implementation of various types of performance tests and to be an Open Source project with the objective of mitigating the use by both the engineering team, QAOPs and Infrastructure.

Its simple format was designed so that tests can be created in minutes and with the possibility of adding them to continuous testing pipelines.

Finally, using the Julia language, it was also created to facilitate the creation of customized metrics. In our project, the following metrics are mapped:

  • Iterations: Total number of requests executed.

  • Success Rate: Percentage of successful requests.

  • Error Rate: Percentage of failed requests.

  • Response Times: Min, Max, Mean, Std, Median, P90, P95, P99.

  • RPS: Requests per second.

  • TPS: Transactions per second.

However, with Julia’s vast mathematical and statistical library, it facilitates the creation of new metrics from the execution return, with a dictionary with all response times.

Installation

To install Stressify, you can use the following command:

$ julia -e 'using Pkg; Pkg.add("Stressify")'

Quick Start

To start using Stressify, you can use the following code:

using Stressify

# Defined the options for the test, in this case we are using 5 VUs, 10 iterations and no duration
Stressify.options(
    vus = 5,
    iterations = 10,
    duration = nothing
)

# After defining the options, we can run the test using the following code
results = Stressify.run_test(
    Stressify.http_get("https://httpbin.org/get"),
    Stressify.http_post("https://httpbin.org/post"; payload="{\"key\": \"value\"}", headers=Dict("Content-Type" => "application/json")),
    Stressify.http_put("https://httpbin.org/put"; payload="{\"update\": \"data\"}", headers=Dict("Content-Type" => "application/json")),
    Stressify.http_patch("https://httpbin.org/patch"; payload="{\"patch\": \"data\"}", headers=Dict("Content-Type" => "application/json")),
    Stressify.http_delete("https://httpbin.org/delete")
)

The code above will run a test with 5 VUs, 10 iterations and no duration. The test will execute 10 iterations of each request, totaling 50 iterations.

The output will be a dictionary with the following structure:

================== Stressify ==================
VUs                    :          5
Iterações Totais       :         50
Taxa de Sucesso (%)    :      100.0
Taxa de Erros (%)      :        0.0
Requisições por Segundo:      10.55
Transações por Segundo :      10.55
Número de Erros        :          0
---------- Métricas de Tempo (s) ----------
Tempo Mínimo           :     0.1609
Tempo Máximo           :     1.2139
Tempo Médio            :     0.4026
Mediana                :     0.3717
P90                    :     0.7145
P95                    :      0.793
P99                    :     1.2139
Desvio Padrão          :     0.2173
---------- Detalhamento de Tempos ----------
Todos os Tempos (s)    : 1.2139, 0.3449, 0.552, 0.3514, 0.2351, 0.1809, 0.3881, 0.3949, 0.3975, 0.1811, 0.7704, 0.6761, 0.3609, 0.3747, 0.1741, 0.2778, 0.3338, 0.45, 0.3367, 0.1683, 0.8838, 0.3846, 0.3703, 0.5831, 0.1645, 0.4906, 0.3726, 0.3708, 0.3667, 0.1713, 0.793, 0.388, 0.5088, 0.3749, 0.172, 0.1714, 0.3786, 0.3254, 0.369, 0.1609, 0.793, 0.5632, 0.3333, 0.3912, 0.1777, 0.1844, 0.4023, 0.4266, 0.7145, 0.1788
==========================================================
---------- Resultados dos Checks ----------

Using Docker to execute Stressify

To run Stressify without installing anything locally, use Docker:

docker pull jfilhogn/stressify:latest
docker run --rm jfilhogn/stressify:latest script.jl

Running the docker image this way will execute the script you want to run, making it easier to execute. In the directory ./docker you can find the Dockerfile and the script.jl file that is used to run the image as example.