Write tailor-made scenarios for load testing web apps
Many good tools already exist for a very long time such as ApacheBench and Siege. The goal is not to replace them. But sometimes, load testing a web app requires to write a custom scenario. My initial requirement was to send requests with unique parameters. To the best of my knowledge, no tool could do this.
The goal is to focus only on the scenario without thinking about forking, threads and non blocking IOs. Fortunately there is Typhoeus to send parallel HTTP requests.
I wrote a blog post where I share my load testing habits. If you are here, that might interest you.
Web Tsunami is a tiny class that forks every seconds and sends as many requests as expected.
It provide the methods get, post, put, patch and delete.
They all accept the same arguments : get(url, options = {}, &block).
The options is given to Typhoeus as is.
It can contain headers and the request body.
See Typhoeus usage for more details.
Let's start with a very trivial scenario and I will show you an advanced one after :
require "web_tsunami"
class SearchTsunami < WebTsunami::Scenario
def run
get("http://site.example") do
# Block is executed once the response has been received
sleep(5) # Simulates the time required for a human to visit the next page
get("http://site.example/search?query=stress+test") do |response|
# Do whatever you need with the response object or ignore it
sleep(10)
get("http://site.example/search?query=stress+test&page=#{rand(100)}") do
sleep(5)
get("http://site.example/stress/test")
end
end
end
end
end
# Simulates 100 concurrent visitors every second for 10 minutes
# It's a total of 60K unique visitors for an average of 23'220 rpm.
SearchTsunami.start(concurrency: 100, duration: 60 * 10)In this scenario, a visitor comes on the index page, then search for stress test, then go on a random page of the search result, and finally found the stress test page. It introduces a unique parameters which is the page number. It's nice, but it could have almost be done with Siege. Let me show you a more realistic scenario.
require "web_tsunami"
class SessionTsunami < WebTsunami::Scenario
def run
# The session object stores cookies and automatically submit CSRF token with forms
session = WebTsunami::Session.new(self, "https://site.example")
session.get("/") do
session.get("/account/new") do
# An authenticity_token param is automatically added by the session
session.post("/account", body: {account: {email: "#{rand(1000000)}@email.test", password: "password"}}) do |response|
# The session stores the Set-Cookie header and will provide it to the next requests
session.get("/dashboard") do # Redirection after registration
# And so on
end
end
end
end
end
end
SessionTsunami.start(concurrency: 100, duration: 60 * 10)This is more realistic because it handles CSRF tokens and cookies. Thus the scenario can submit forms and behaves a little bit more like a real visitor.
Web Tsunami does not measure response time or print any result. If you are running a load test, that means you should use an APM (application performance monitoring). It already collects and displays all data you need such as throughput, response time and bottlenecks.
So, the only output are errors.
The gem is available as open-source under the terms of the MIT License.
Rails developer? Check out RoRvsWild, our Ruby on Rails application monitoring tool.