A performance testing framework for cucumber io.
Cucumber perf represents a new idea in automated testing development.
Cucumber is a implementation of Behavior Driven Development (BDD). Which uses simple natural language scripts to define a software feature. These executable specifications are written in a language called gherkin. Example:
Feature: Beer
Scenario: Jeff drinks a beer
Given: Jeff is of age and has a beer
And: Jeff opens his beer.
When: Jeff takes a sip.
Then: Verify he enjoyed it.
These scripts can be used to develop the features themselves but also drive automated tests.
So, you now have a working functional automation test suite. But you want to run a performance test. This would require either rewriting your existing functional tests or copying a bunch of code. Also, you would need to create a performance test harness.
Most likely each team will end up with something that is project specific and doesn't use the existing functional code base.
Cucumber Perf provides a level of automation on top of Cucumber. It’s an implementation of a new concept (as far as I know) called Concurrent Behavior Driven Testing (CBDT).
Cucumber perf provides a means to use your existing functional code without writing a single line of code. It provides the ability to run performance simulations with support for common load testing features:
- Timed Tests
- Multi-Threading
- Thread Count Limits
- Ramp Up/Down
- Data replacing
- Random Wait
- Reporting
- Logging
It uses a type of script called Salad. Salad is a reimplementation of Cucumber Gherkin with the focus on performance simulations.
Plan: Bar visit
Simulation: Jeff drinks 3 beers.
Group: beer.feature
Runners: 1
Count: 3
Here is an example plan
Plan: test
Simulation: simulation 1
Group test.feature
#slices
#these values will replace property "value out"
|value out|
|changed value 1|
|changed value 2|
#number of threads
Runners: 2
#total number of threads to run.
Count: 2
#a optional random wait mean for before thread runs tests.
#thread will wait between +-50% of this mean
RandomWait: 00:00:02
#Will run all groups for the period below
Simulation Period: simulation 2 period
Group test.feature
|value out|
|changed value |
Threads: 5
#count is ignored in a simulation period
Count: 1
#run time
Time: 00:00:30
RampUp: 00:00:10
RampDown: 00:00:10
Concurrent Behavior Driven Testing is the methodology of creating functional automation that can be used in concurrent test scenarios. This means coding with the understanding that each functional test could be used in a multithreaded environment.
CBDT requires an automation team to follow strict guidelines when coding functional test cases. Being careful to avoid static variables and race conditions that will cause failures in a multi-threaded world. This of course requires a larger understanding of programing or a least team leadership that can enforce these guidelines.
It takes planning to implement Cucumber Perf.
Your functional automation should follow these rules:
- Use a non specific test harness. This should standardize all your common functions.
- Do not use static variables! Your code must work in a multithreaded world.
- Properly comment your features and scenarios. You want to keep track of what scenarios can be run multithreaded.
Follow directions in wiki to get up and running.
We use SemVer for versioning. For the versions available, see the tags on this repository.
This project is licensed under the MIT License - see the LICENSE.md file for details