Thank you for your interest in contributing to k6!
(ノ◕ヮ◕)ノ*:・゚✧
Before you begin, make sure to familiarize yourself with the Code of Conduct. If you've previously contributed to other open source project, you may recognize it as the classic Contributor Covenant.
If you want to chat with the team or the community, you can join our community forums.
Note: To disclose security issues, refer to SECURITY.md.
Don't be afraid to file issues! Nobody can fix a bug we don't know exists, or add a feature we didn't think of.
The worst that can happen is that someone closes it and points you in the right direction.
That said, "how do I..."-type questions are often more suited for community forums.
If you'd like to contribute code to k6, this is the basic procedure. Make sure to follow the style guide described below.
-
Find an issue you'd like to fix. If there is none already, or you'd like to add a feature, please open one, and we can talk about how to do it. Out of respect for your time, please start a discussion regarding any bigger contributions either in a GitHub Issue, in the community forums before you get started on the implementation.
Remember, there's more to software development than code; if it's not properly planned, stuff gets messy real fast.
-
Create a fork and open a feature branch -
feature/my-cool-feature
is the classic way to name these, but it really doesn't matter. -
Create a pull request!
-
Sign the Contributor License Agreement (the process is integrated with the pull request flow through cla-assistant.io).
-
We will discuss implementation details until everyone is happy, then a maintainer will merge it.
To get a basic development environment for Go and k6 up and running, first make sure you have Git and Go (see our go.mod for minimum required version) installed and working properly.
We recommend using the Git command-line interface to download the source code for the k6:
- Open a terminal and run
git clone https://github.com/grafana/k6.git
. This command downloads k6's sources to a newk6
directory in your current directory. - Open the
k6
directory in your favorite code editor.
For alternative ways of cloning the k6 repository, please refer to GitHub's cloning a repository documentation.
We make use of the golangci-lint tool to lint the code in CI. The actual version you can find in our .golangci.yml
. To run it locally, first install it, then run:
make lint
To exercise the entire test suite, please run the following command
make tests
Consult Dependencies.md to find out more about how we manage k6's dependencies, and our policy regarding dependencies management and update.
As you'd expect, please adhere to good ol' gofmt
(there are plugins for most editors that can autocorrect this), but also gofmt -s
(code simplification), and don't leave unused functions laying around.
Continuous integration will catch all of this if you don't, and it's fine to just fix linter complaints with another commit, but you can also run the linter yourself:
make check
Comments in the source should wrap at 100 characters, but there's no maximum length or need to be brief here - please include anything one might need to know in order to understand the code, that you could reasonably expect any reader to not already know (you probably don't need to explain what a goroutine is).
We don't have any explicit rules about commit message formatting, but try to write something that could be included as-is in a changelog.
If your commit closes an issue, please close it with your commit message, for example:
Added this really rad feature
Closes #420
Any human-readable text you add must be non-gendered and should be fairly concise without devolving into grammatical horrors, dropped words, and shorthands. This isn't Twitter, you don't have a character cap, but don't write a novel where a single sentence would suffice.
If you're writing a longer block of text to a terminal, wrap it at 80 characters - this ensures it will display properly at the de facto default terminal size of 80x25. As an example, this is the help text of the scale
sub-command:
Scale will change the number of active VUs of a running test.
It is an error to scale a test beyond vus-max; this is because instantiating
new VUs is a very expensive operation, which may skew test results if done
during a running test. To raise vus-max, use --max/-m.
Endpoint: /v1/status