Contributing¶

Thank you for wanting to contribute to gcp-nuke.

Because of the amount of gcp services and their rate of change, we rely on your participation. For the same reason we can only act retroactive on changes of gcp services. Otherwise, it would be a full time job to keep up with gcp.

How Can I Contribute?¶

Some Resource Is Not Supported by gcp-nuke¶

If a resource is not yet supported by gcp-nuke, you have two options to resolve this:

File an issue and describe which resource is missing. This way someone can take care of it.
Add the resource yourself and open a Pull Request. Please follow the guidelines below to see how to create such a resource.

Some Resource Does Not Get Deleted¶

Please check the following points before creating a bug issue:

Is the resource actually supported by gcp-nuke? If not, please follow the guidelines above.
Are there permission problems? In this case gcp-nuke will print errors that usually contain the status code 403.
Did you just get scared by an error that was printed? gcp-nuke does not know about dependencies between resources. To work around this it will just retry deleting all resources in multiple iterations. Therefore, it is normal that there are a lot of dependency errors in the first one. The iterations are separated by lines starting with Removal requested: and only the errors in the last block indicate actual errors.

File an issue and describe as accurately as possible how to generate the resource on gcp that cause the errors in gcp-nuke. Ideally this is provided in a reproducible way like a Terraform template or gcp CLI commands.

I Have Ideas to Improve gcp-nuke¶

You should take these steps if you have an idea how to improve gcp-nuke:

Check the issues page, whether someone already had the same or a similar idea.
Also check the closed issues, because this might have already been implemented, but not yet released. Also, the idea might not be viable for obvious reasons.
Join the discussion, if there is already a related issue. If this is not the case, open a new issue and describe your idea. Afterward, we can discuss this idea and form a proposal.

I Just Have a Question¶

Please use GitHub Discussions

Resource Guidelines¶

Tooling¶

Checkout the documentation around resources as it provides resource format and a tool to help generate the resource.

Consider Pagination¶

Most gcp resources are paginated and all resources should handle that.

Use Properties Instead of String Functions¶

Currently, each resource can offer two functions to describe itself, that are used by the user to identify it and by gcp-nuke to filter it.

The String function is deprecated:

func (r *Resource) String() string

The Properties function should be used instead:

func (r *Resource) Properties() types.Properties

Note: The interface for the String function is still there, because not all resources are migrated yet. Please use the Properties function for new resources.

Filter Resources That Cannot Get Removed¶

Some gcp APIs list resources, that cannot be deleted. For example:

Resources that are already deleted, but still listed for some time (e.g. EC2 Instances)
Resources that are created by gcp, but cannot be deleted by the user (e.g. some IAM Roles)

Those resources should be excluded in the filter step, rather than in the list step.

Styleguide¶

Go¶

golangci-lint¶

There is an extensive golangci-lint configuration in the repository. Please make sure to run golangci-lint run before committing any changes.

Code Format¶

Like almost all Go projects, we are using go fmt as a single source of truth for formatting the source code. Please use go fmt before committing any change.

Import Format¶

Standard library imports
Third party imports
gcp SDK imports
ekristen/libnuke imports
Local package imports

Example Import Format¶

package example

import (
    "context"

    "github.com/sirupsen/logrus"

    "github.com/gcp/gcp-sdk-go-v2/gcp"
    "github.com/gcp/gcp-sdk-go-v2/service/s3"

    "github.com/ekristen/libnuke/pkg/settings"

    "github.com/ekristen/gcp-nuke/pkg/types"
)

Git¶

Pull Requests¶

We default to squash merge pull requests. This means that the commit history of the pull request will be squashed into a single commit and then merged into the main branch. This keeps the commit history clean and easy to read.

Commits¶

We are using the Conventional Commits specification for our commit messages. This allows us to automatically generate a changelog and version numbers.

All commits in a pull request must follow this format or the GitHub Actions will fail.

Signed Commits¶

We require that all commits be signed.

git config --global commit.gpgsign true

Setup Email¶

We prefer having the commit linked to the GitHub account, that is creating the Pull Request. To make this happen, git must be configured with an email, that is registered with a GitHub account.

To set the email for all git commits, you can use this command:

git config --global user.email "email@example.com"

If you want to change the email only for the gcp-nuke repository, you can skip the --global flag. You have to make sure that you are executing this in the gcp-nuke directory:

git config user.email "email@example.com"

If you already committed something with a wrong email, you can use this command:

git commit --amend --author="Author Name <email@address.com>"

This changes the email of the latest commit. If you have multiple commits in your branch, please squash them and change the author afterwards.