How to contribute to Apache NiFi
We are always excited to have contributions from the community - especially from new contributors!
We are interested in accepting contributions of code, as well as documentation and even artwork that can be applied as icons or styling to the application.
Documentation is created in AsciiDoc.
Running NiFi in Debug mode
While NiFi Mock library can greatly simplify your development and testing efforts, some times it is still necessary to be able to set breakpoints and debug your code within the running instance of NiFi. To do that NiFi needs to be started in Debug mode within your IDE. Please follow the instructions provided in the NiFi IDE Integration module.
Where to Start?
Finding Issues and Extending the Project
NiFi's JIRA page can be used to find tickets that are tagged as "beginner", or you can dig into any of the tickets for creating Processors. Processors should be self-contained and not rely on other outside components (except for Controller Services), so they make for excellent starting points for new NiFi developers to get started. This exposes the developer to the NiFi API and is the most extensible part of the dataflow system. The NiFi Developer Guide provides details on how to get started with the NiFi API as well as best practices and common approaches to component design.
We have additionally created filters that show various types of issues available to be worked:
System-level and overview documentation such as:
- Administration Guide,
- Developer's Guide,
- User's Guide,
Are located in
'<code checkout location>/nifi-docs/src/main/asciidoc' and can be changed by making changes to the ASCIIdoc files and following the code contribution process described in "Providing code or documentation contributions" section of this document.
Tools available to facilitate documentation generation are available at Editing AsciiDoc with Live Preview.
Component level documentation can be contributed to by making direct modifications to the source code of the relevant component. Generally the documentation of components is performed by adding the relevant details to the following annotations
and the description field of annotations such as:
and the description method of Component Properties and Relationships
When the documentation of a processor is particularly complex (e.g, may include very long descriptions or images), a contributor may choose to use NiFi's capability to automatically link additional documentation to a processor.
In this case the contributor may extend documentation by performing changes to
'additionalDetails.html'. As described in the Developer's Guide, this file should exist within a directory whose name is the fully-qualified name of the Processor, and this directory’s parent should be named docs and exist in the
'resources' directory of a particular processor bundle.
As with other contributions, component level documentation should follow the process described in "Providing code or documentation contributions" section of this document.
Create a ticket for new bugs or features
Run into a bug or think there is something that would benefit the project? Regardless if you have the time to provide the fix or implementation, we encourage any such items to be filed as an issue at the Apache NiFi JIRA.
Providing code or documentation contributions
The source is hosted at
Configure your git client
Ensure your git user name and email are configured
The following lines ensure your commits are appropriately annotated with your information
git config --global user.name "User Name"
git config --global user.email email@example.com
Windows Specific configuration
The following options provide handling of long file paths that can be troublesome as well as not using Windows style line returns.
git config --global core.longpaths true
git config --global core.autocrlf false
Clone a copy of the repository
From the Apache Hosted Repository
From the GitHub Mirror
git clone https://github.com/apache/nifi.git
From your GitHub Fork
git clone firstname.lastname@example.org:<account name>/nifi.git
Retrieval of upstream changes
Additionally, it is beneficial to add a git remote for the mirror to allow the retrieval of upstream changes
git remote add upstream https://github.com/apache/nifi.git
Checkout the 'master' or '1.x.0' branch
git checkout -b master origin/master will create a local, tracking branch named
git checkout -b 0.x origin/1.x.0 will create a local, tracking branch named 1
The master branch currently represents the next major release line (1.x.0).
The NiFi community uses a modified Gitflow development model. A summary:
- Use of a central repository
- Branch per feature similar to the Feature Branch Workflow
- Work is done locally and then pushed to the central repo
- 'master' branch contains the official release history. Code changes (not code formatting, administrative updates) require Review-Then-Commit (RTC) by another committer to get incorporated.
Perform your code changes
Create a feature branch
Typically, there is a 1-to-1 mapping of a branch to a backing ticket as specified in JIRA.
Create a local branch that relates the associated JIRA issue with the branch. Such an example would be:
git checkout -b nifi-359 master
This provides instant traceability to the supporting issue and provides a means of linking discussion.
Test your changes
For code changes, ensure that the full suite of tests is executed via
mvn -Pcontrib-check clean install at the root nifi folder. Please write or update unit tests to verify your changes.
For documentation related changes, ensure that format looks appropriate for the output in which it is rendered.
Update Licensing Documentation
Did your change introduce new dependencies? Are these dependencies licensed in a way that is compatible for inclusion under ASF 2.0? Did you change or remove any existing dependency? Did you update the LICENSE and NOTICE files?
Each new source or configuration file added must have the ASF 2.0 license at the beginning of the file. It should be easy enough to copy one from another location.
There are a few places where licensing matters when you're making software changes. There are global LICENSE and NOTICE files in the nifi-assembly module which rolls up the contents of each NIFI nar LICENSE and NOTICE. The general guidelines for maintaining these are as follows:
- If you've added dependencies, you need to ensure that the dependencies are licensed in a way that is compatible for inclusion under ASF 2.0. These are generally defined here.
- If you've changed a dependency at all (e.g. add, remove, change version), you must verify that the LICENSE and NOTICE from that dependency are contained the nar you're changing (if applicable) and in nifi-assembly. Be mindful to check if the dependency you're adding already exists in the LICENSE and NOTICE files you're editing. If it does, Consider the following guidelines:
- LICENSE: If the dependency version already exists here and the licenses are identical, you may simply add the additional version (e.g. "The binary distribution of this product bundles 'Google Protocol Buffers Java 2.5.0 and 3.3.1"). If the licenses are different, list them separately.
- NOTICE: If the dependency version already exists here and the notices are the same, there's no change needed. If the notices are different you need to determine how different they are. If one is a superset of the other, you may simply replace the subset with the superset. If they're vastly different from each other, you need to list them separately.
Other general guidelines as they relate to licensing in NIFI:
- Whether some binary dependencies NOTICE calls out a transitive binary dependency it might or might not have is not relevant. What is relevant is which transitive dependencies, no matter how many levels deep it comes in, we pull into our nars or convenience binaries. They must all be accounted for properly if we're including them. See the Apache Licensing Guide for more information on this.
- All binary dependencies bundled in a given nar must be documented in its LICENSE and NOTICE. On top of that, the nifi-assembly should cover those dependencies as well.
- All source dependencies taken as-in or modified must be documented in the nifi-assembly LICENSE and NOTICE.
Commit your changes
List the current files that have changes
$ git status
On branch nifi-359
Changes not staged for commit:
(use "git add <file>..." to update what will be committed)
(use "git checkout -- <file>..." to discard changes in working directory)
no changes added to commit (use "git add" and/or "git commit -a")
Stage the file(s) to be committed with
$ git add nifi-docs/src/main/asciidoc/contributor-guide.adoc
$ git status
On branch nifi-359
Changes to be committed:
(use "git reset HEAD <file>..." to unstage)
In the case of multiple files, it is also possible to stage all tracked files through
git add .
Perform the commit
In the interest of providing traceability, it is helpful to lead off your commit with the associated JIRA issue number (case-sensitive, NIFI in uppercase) and a summary of the change this commit is providing. Such an example is:
$ git commit -m "NIFI-359 Adding a guide for procedure and best practices in contributing to the project."
The issue being listed first provides easy access to the supporting ticket and ensures it is not truncated from the varied means in which commits will be viewed.
Keeping your feature branch current
If you are working on a branch over an extended period of time, it helps to keep your code current with the master branch to ensure your changes are applied as anticipated and to facilitate the merging process.
This is best accomplished through the
git rebase functionality. There are many ways in which
git rebase can be utilized in this context of branching and rebasing. Typically, the command to do so is performed from your feature branch. Continuing on with the sample of NIFI-359, we will show one way of accomplishing this task.
Update your local copy of
$ git checkout master
Switched to branch 'master'
$ git fetch upstream
remote: Counting objects: 52, done.
remote: Compressing objects: 100% (11/11), done.
remote: Total 16 (delta 7), reused 0 (delta 0)
Unpacking objects: 100% (16/16), done.
a49a03d..8d745c2 master -> upstream/master
$ git merge upstream/master
nifi-nar-bundles/nifi-framework-bundle/nifi-framework/nifi-framework-cluster-protocol/src/main/java/org/apache/nifi/cluster/protocol/StandardDataFlow.java | 27 ++++++++++-----------------
1 file changed, 10 insertions(+), 17 deletions(-)
git rebase from your branch
$ git checkout nifi-359
Switched to branch 'nifi-359'
$ git rebase master
First, rewinding head to replay your work on top of it...
Applying: NIFI-359: Removing the output to System.out and providing some simple checks to ensure extraction of attributes is consistent with the sample file used in the test.
Supplying a contribution
There are currently two paths from which we review and accept contributions, ASF JIRA and GitHub Pull Requests (PRs), regardless of whether your contribution is source code or documentation.
Supplying a contribution through patch and JIRA issue
Assuming you are still on the branch you want to make a patch from (NIFI-359) and that you've just committed the change and it is a single commit then you can run the following to generate a patch file
`git format-patch nifi-359~1 -o </desired/path/for/patch>/nifi-359.patch`
Now you can go to JIRA for ticket NIFI-359 and select 'more -> attach files' and upload the patch file. Then select 'Submit Patch'.
TODO: Once auto-build verification comes into play this guidance will likely change.
Supplying a contribution through a pull request (PR) via GitHub
Push changes to your personal, GitHub repository remote
Typically, this remote is listed as
origin. To confirm, perform a
git remote -v from within your cloned repository directory.
$ git remote -v
origin email@example.com:apiri/nifi.git (push)
origin firstname.lastname@example.org:apiri/nifi.git (fetch)
upstream https://git-wip-us.apache.org/repos/asf/nifi.git (fetch)
Once the appropriate remote (remote name) is determined, it is possible to push the branch with your changes to GitHub through:
$ git push <remote name> <branch name>
As an example, based on the context of this guide, this would take the form of:
$ git push origin nifi-359
This will update your fork of NiFi and can be verified from your GitHub personal project page:
Opening a Pull Request (PR) to the NiFi project
From your personal fork, a Pull Request can be opened simply by selecting the Pull Request link highlighted below.
This will allow specification of the your branch which has the changes, and will open a PR for review by the committers of the project.
Code Review Process
Apache NIFI has a Review-Then-Commit (RTC) philosophy for handling all contributions. Reviewers first ensure that the code applies and builds appropriately to the build, passes the code standards as established by the Maven profile "contrib-check." From here, code is evaluated to ensure best practices within the NiFi framework are applied and, where applicable, that the user experience of interfacing with the contribution is consistent and any changes are backwards compatible. This process may be iterative but works toward a final product that is then merged into the codebase.
While only committers can actively promote contributions into the repository, feedback on issues, regardless of committer status, is appreciated and valued in the review process.
If you are interested in facilitating the review process, a listing of all code contributions with a patch are available via a JIRA filter, NIFI Patch Available.
A view of the filter is shown below:
Additional notes on code contributions
Rebasing and Squashing while under review
Although you may be asked to rebase or squash your contribution as part of the review process, don't feel the need to do so speculatively. The committer working on merging the contribution may prefer to do these types of operations as part of the merge process, and the history of your patch or pull request may aid in the review process.
Steps to merge/close pull requests with two main branches
As NiFi now has a 1.0 (master) and 0.x (support) branch, pull requests (PR) must be applied to both. Here is a step-by-step guide for committers to ensure this occurs for all PRs.
- Check out the latest master
$ git checkout master
- $ git pull upstream master
- Check out the PR (example #327). This will be in
detached-HEADstate. (Note: You may need to edit the
.git/configfile to add the
$ git checkout github/pr/327
- Create a branch for the PR
- $ git checkout -b pr327
- Apply the changes and sign off. This could be through a
$ git commit --amend -s
- Edit the commit file to contain "This closes #327. "
$ git log
- Copy commit id of last commit
- Switch back to the master branch
$ git checkout master
- Merge the changes. You can use
$ git cherry-pick <commit id>
- (Optional) Ensure the commit was applied successfully
$ git log
- Push to the Apache repository (master branch)
$ git push apache master
- Switch to the support branch
$ git checkout -t upstream/0.x
- (Optional) Check the status of the branch
$ git log
- Apply the changes from the PR branch
$ git cherry-pick <commit id>
- Push to the Apache repository (support branch)
$ git push apache 0.x
To ensure you are able to pull the PR directly, add the following lines to your
url = email@example.com:apache/nifi.git
fetch = +refs/heads/*:refs/remotes/github/*
fetch = +refs/pull/*/head:refs/remotes/github/pr/*
TODO: Provide code formatters for the common IDEs
IntelliJ IDEA Users
nifi-checkstyle.xml is a Checkstyle configuration file created by extracting
Steps for IntelliJ users:
- Download Checkstyle-IDEA plugin
- Install plugin by loading the downloaded zip from:
Install plugin from disk...
- After restarting IntelliJ, configure Checkstyle plugin to use NiFi Checkstyle configuration file from:
nifi-checkstyle.xmlin Configuration File and activate it
You can now see Checkstyle violations in the editor in real-time.
Code formatter rules that adhere to our checkstyle rules can be found here.
Consider installing these templates that help generate common Apache NiFi code snippets. TODO: provide link
The developer mailing list (firstname.lastname@example.org) is monitored pretty closely, and we tend to respond quickly. If you have a question, don't hesitate to shoot us an e-mail - we're here to help! Unfortunately, though, e-mails can get lost in the shuffle, so if you do send an e-mail and don't get a response within a day or two, it's our fault - don't worry about bothering us. Just ping the mailing list again.