wrslatz/bulldozer

GitHub Pull Request Auto-Merge Bot

bulldozer

bulldozer is a GitHub App that automatically merges pull requests (PRs) when (and only when) all required status checks are successful and required reviews are provided.

Additionally, bulldozer can:

• Only merge pull requests that match certain conditions, like having a specific label or comment
• Ignore pull requests that match certain conditions, like having a specific label or comment
• Automatically keep pull request branches up-to-date by merging in the target branch
• Wait for additional status checks that are not required by GitHub

Bulldozer might be useful if you:

• Have CI builds that take longer than the normal review process. It will merge reviewed PRs as soon as the tests pass so you don't have to watch the pull request or remember to merge it later.
• Combine it with policy-bot to automatically merge certain types of pre-approved or automated changes.
• Want to give contributors more control over when they can merge PRs without granting them write access to the repository.
• Have a lot of active development that makes it difficult to merge a pull request while it is up-to-date with the target branch.

Contents

• Behavior
• Configuration
  • bulldozer.yml Specification
• FAQ
• Deployment
• Development
• Contributing
• License

Behavior

bulldozer will only merge pull requests that GitHub allows non-admin collaborators to merge. This means that all branch protection settings, including required status checks and required reviews, are respected. It also means that you must enable branch protection to prevent bulldozer from immediately merging every pull request.

Only pull requests matching the trigger conditions (or not matching ignore conditions) are considered for merging. bulldozer is event-driven, which means it will usually merge a pull request within a few seconds of the pull request satisfying all preconditions.

Configuration

The behavior of the bot is configured by a .bulldozer.yml file at the root of the repository. The file name and location are configurable when running your own instance of the server.

• The file is read from the most recent commit on the target branch of each pull request.

• The file may contain a reference to a configuration in a different repository (see Remote Configuration.)

• If the file does not exist in the repository, bulldozer tries to load a shared bulldozer.yml file at the root of the .github repository in the same organization. You can change this path and repository name when running your own instance of the server.

• You can also define a global default configuration when running your own instance of the server. This is used if the repository or the shared organization repository do not define any configuration.

• If configuration does not exist in the repository or in the shared organization repository and the server does not have a default configuration, bulldozer does not act on the pull request. This means it is safe to enable bulldozer on all repositories in an organization.

bulldozer.yml Specification

The .bulldozer.yml file supports the following keys.

# "version" is the configuration version, currently "1".
version: 1

# "merge" defines how and when pull requests are merged. If the section is
# missing, bulldozer will consider all pull requests and use default settings.
merge:
  # "trigger" defines the set of pull requests considered by bulldozer. If
  # the section is missing, bulldozer considers all pull requests not excluded
  # by the ignore conditions.
  trigger:
    # Pull requests with any of these labels (case-insensitive) are added to
    # the trigger.
    labels: ["merge when ready"]

    # Pull requests where the body or any comment contains any of these
    # substrings are added to the trigger.
    comment_substrings: ["==MERGE_WHEN_READY=="]

    # Pull requests where any comment matches one of these exact strings are
    # added to the trigger.
    comments: ["Please merge this pull request!"]

    # Pull requests where the body contains any of these substrings are added
    # to the trigger.
    pr_body_substrings: ["==MERGE_WHEN_READY=="]

    # Pull requests targeting any of these branches are added to the trigger.
    branches: ["develop"]

    # Pull requests targeting branches matching any of these regular expressions are added to the trigger.
    branch_patterns: ["feature/.*"]

    # Pull requests with auto merge enabled are added to the trigger.
    auto_merge: true

  # "ignore" defines the set of pull request ignored by bulldozer. If the
  # section is missing, bulldozer considers all pull requests. It takes the
  # same keys as the "trigger" section.
  ignore:
    labels: ["do not merge"]
    comment_substrings: ["==DO_NOT_MERGE=="]

  # "method" defines the merge method. The available options are "merge",
  # "rebase", "squash", and "ff-only".
  method: squash

  ##### branch_method has been DEPRECATED in favor of merge_method #####
  # 
  # Allows the merge method that is used when auto-merging a PR to be different
  # target branch. The keys of the hash are the target branch name, and the values are the merge method that
  # will be used for PRs targeting that branch. The valid values are the same as for the "method" key.
  # Note: If the target branch does not match any of the specified keys, the "method" key is used instead.
  branch_method:
    develop: squash
    master: merge
  ##### branch_method has been DEPRECATED in favor of merge_method #####

  # Allows the merge method that is used when auto-merging a PR to be different
  # based on trigger criteria. The first method where ALL triggers match will
  # be used. Otherwise, the method specified previously in "merge.method" will 
  # be used.
  # - ALL trigger criteria must match, unlike merge/trigger where ANY match 
  # will trigger bulldozer.
  # - This will override any branch_method logic if one of the methods is
  # triggered
  # - If no trigger criteria is provided the method is ignored
  merge_method:
    # "method" defines the merge method. The available options are "merge",
    # "rebase", "squash", and "ff-only".
    - method: squash
      trigger:
        # All methods from merge/trigger are supported. Additionally, the
        # following additional methods are provided:

        # Pull requests which a number of commits less than or equal to this value are added to the trigger.
        max_commits: 3

  # "options" defines additional options for the individual merge methods.
  options:
    # "squash" options are only used when the merge method is "squash"
    squash:
      # "title" defines how the title of the commit message is created when
      # generating a squash commit. The options are "pull_request_title",
      # "first_commit_title", and "github_default_title". The default is
      # "pull_request_title".
      title: "pull_request_title"

      # "body" defines how the body of the commit message is created when
      # generating a squash commit. The options are "pull_request_body",
      # "summarize_commits", and "empty_body". The default is "empty_body".
      body: "empty_body"

      # If "body" is "pull_request_body", then the commit message will be the
      # part of the pull request body surrounded by "message_delimiter"
      # strings. This is disabled (empty string) by default.
      message_delimiter: ==COMMIT_MSG==

  # "required_statuses" is a list of additional status contexts that must pass
  # before bulldozer can merge a pull request. This is useful if you want to
  # require extra testing for automated merges, but not for manual merges.
  required_statuses:
    - "ci/circleci: ete-tests"

  # If true, bulldozer will delete branches after their pull requests merge.
  delete_after_merge: true

  # If true, bulldozer will merge pull requests with no required checks. This
  # helps to protect against merging branches which inadvertently do not have
  # required status checks.
  allow_merge_with_no_checks: false

# "update" defines how and when to update pull request branches. Unlike with
# merges, if this section is missing, bulldozer will not update any pull requests.
update:
  # "trigger" defines the set of pull requests that should be updated by
  # bulldozer. It accepts the same keys as the trigger in the "merge" block.
  trigger:
    labels: ["WIP", "Update Me"]

  # "ignore" defines the set of pull requests that should not be updated by
  # bulldozer. It accepts the same keys as the ignore in the "merge" block.
  ignore:
    labels: ["Do Not Update"]

  # If true, bulldozer will ignore updating draft pull requests, unless they
  # explicitly match a configured trigger condition.
  ignore_drafts: false

Remote Configuration

You can use a remote configuration by specifying a repository and an optional path and Git reference. Place the following in the repository's .bulldozer.yml file instead of the normal configuration:

# The remote repository to read the configuration file from. This is required,
# and must be in "org/repo-name" form. Must be a public repository.
remote: org/repo-name

# The path to the configuration file in the remote repository. If not set,
# uses the default configuration path.
path: path/to/bulldozer.yml

# The branch (or tag, or commit hash) that should be used on the remote
# repository. If not set, uses the default branch of the repository.
ref: main

The remote file must contain bulldozer configuration and cannot be another remote reference. However, the organization-level default configuration may be a remote reference.

FAQ

Can I specify both ignore and trigger?

Yes. If both ignore and trigger are specified, bulldozer will attempt to match on both. In cases where both match, ignore will take precedence.

Can I specify the body of the commit when using the squash strategy?

Yes. When the merge strategy is squash, you can set additional options under the options.squash property, including how to render the commit body.

merge:
  method: squash
  options:
    squash:
      body: summarize_commits # or `pull_request_body`, `empty_body`

You can also define part of pull request body to pick as a commit message when body is pull_request_body.

merge:
  method: squash
  options:
    squash:
      body: pull_request_body
      message_delimiter: ==COMMIT_MSG==

Anything that's contained between two ==COMMIT_MSG== strings will become the commit message instead of whole pull request body.

What if I don't want to put config files into each repo?

You can add default repository configuration in your bulldozer config file.

It will be used only when your repo config file does not exist.

options:
  default_repository_config:
    ignore:
      labels: ["do not merge"] # or any other available config.

Bulldozer isn't merging my commit when it should, what could be happening?

Bulldozer will attempt to merge a branch whenever it passes the trigger/ignore criteria. GitHub may prevent it from merging a branch in certain conditions, some of which are to be expected, and others that may be caused by mis-configuring Bulldozer.

• Required status checks have not passed
• Review requirements are not satisfied
• The merge strategy configured in .bulldozer.yml is not allowed by your repository settings
• Branch protection rules are preventing bulldozer[bot] from pushing to the branch. Github apps can be added to the list of restricted push users, so you can allow Bulldozer specifically for your repo.
• The branch has no required checks and allow_merge_with_no_checks is set to the default value (false).

Bulldozer isn't updating my branch when it should, what could be happening?

When using the branch update functionality, Bulldozer only performs an update after updates are enabled when:

• A label is added
• The pull request is opened
• The target branch is updated

For example:

1. User A opens a pull request targetting develop
2. User B pushes a commit to develop
3. User A adds an update me comment to the first pull request
4. User C pushes a commit to develop
5. Bulldozer updates the pull request with the commits from Users B and C

Note that the update does not happen when the update me comment is added, even though there is a new commit on develop that is not part of the pull request.

Can Bulldozer work with push restrictions on branches?

As mentioned above, as of Github ~2.19.x, GitHub Apps can be added to the list of users associated with push restrictions. If you don't want to do this, or if you're running an older version of Github that doesn't support this behaviour, you may work around this:

1. Use another app like policy-bot to implement approval restrictions as required status checks instead of using push restrictions. This effectively limits who can push to a branch by requiring changes to go through the pull request process and be approved.

2. Configure Bulldozer to use a personal access token for a regular user to perform merges in this case. The token must have the repo scope and the user must be allowed to push to the branch. In the server configuration file, set:

   options:
     push_restriction_user_token: <token-value>

   The token is only used if the target branch has push restrictions enabled. All other merges are performed as the normal GitHub App user.

Deployment

bulldozer is easy to deploy in your own environment as it has no dependencies other than GitHub. It is also safe to run multiple instances of the server, making it a good fit for container schedulers like Nomad or Kubernetes.

We provide both a Docker container and a binary distribution of the server:

• Binaries: https://github.com/palantir/bulldozer/releases
• Docker Images: https://hub.docker.com/r/palantirtechnologies/bulldozer/

A sample configuration file is provided at config/bulldozer.example.yml. Certain values may also be set by environment variables; these are noted in the comments in the sample configuration file. By default, the environment variables for server values are prefixed with BULLDOZER_ (e.g. BULLDOZER_PORT). This prefix can be overridden by setting the BULLDOZER_ENV_PREFIX environment variable.

GitHub App Configuration

To configure Bulldozer as a GitHub App, these general options are required:

• Webhook URL: http(s)://<your-bulldozer-domain>/api/github/hook
• Webhook secret: A random string that matches the value of the github.app.webhook_secret property in the server configuration

The app requires these permissions:

Permission	Access	Reason
Repository administration	Read-only	Determine required status checks
Checks	Read-only	Read checks for ref
Repository contents	Read & write	Read configuration, perform merges
Issues	Read & write	Read comments, close linked issues
Repository metadata	Read-only	Basic repository data
Pull requests	Read & write	Merge and close pull requests
Commit status	Read-only	Evaluate pull request status

The app should be subscribed to these events:

• Check run
• Commit comment
• Pull request
• Status
• Push
• Issue comment
• Pull request review
• Pull request review comment

Operations

bulldozer uses go-baseapp and go-githubapp, both of which emit standard metrics and structured log keys. Please see those projects for details.

Example Files

Example .bulldozer.yml files can be found in config/examples

Migrating: Version 0.4.X to 1.X

The server configuration for bulldozer allows you to specify configuration_v0_path, which is a list of paths to check for 0.4.X style bulldozer configuration. When a 1.X style configuration file does not appear at the configured path, bulldozer will attempt to read from the paths configured by configuration_v0_path, converting the legacy configuration into an equivalent v1 configuration internally.

The upgrade process is therefore to deploy the latest version of bulldozer with both configuration_path and configuration_v0_path configured, and to enable the bulldozer GitHub App on all organizations where it was previously installed.

Reducing Update Commit and CI Pressure

The Bulldozer "Update" feature is useful when keeping branches up-to-date, but can cause lots of pressure on CI build systems and Github when there are many pull requests open on a single repository.

To reduce pressure on CI systems and Github, the update feature can be disabled at the server level by specifying the following server option:

options:
  disable_update_feature: true

Development

To develop bulldozer, you will need a Go installation.

Run style checks and tests

./godelw verify

Running the server locally

# copy and edit the server config
cp config/bulldozer.example.yml config/bulldozer.yml

./godelw run bulldozer server

• config/bulldozer.yml is used as the default configuration file
• The server is available at http://localhost:8080/

Running the server via Docker

# copy and edit the server config
cp config/bulldozer.example.yml config/bulldozer.yml

# build the docker image
./godelw docker build --verbose

docker run --rm -v "$(pwd)/config:/secrets/" -p 8080:8080 palantirtechnologies/bulldozer:latest

• This mounts the config directory (which should contain the bulldozer.yml configuration file) at the expected location
• The server is available at http://localhost:8080/

Contributing

Contributions and issues are welcome. For new features or large contributions, we prefer discussing the proposed change on a GitHub issue prior to a PR.

License

This application is made available under the Apache 2.0 License.