A light-weight lint workflow to unite the worlds of GitHub and Jira. Based on the fine work of ClearTax on jira-lint.
To make action-jira-linter
a part of your workflow, just add a
action-jira-linter.yml
file in your .github/workflows/
directory in your
GitHub repository.
name: action-jira-linter
on: [pull_request]
jobs:
action-jira-linter:
runs-on: ubuntu-latest
steps:
- uses: btwrk/[email protected]
with:
github-token: ${{ secrets.GITHUB_TOKEN }}
jira-token: ${{ secrets.JIRA_TOKEN }}
jira-base-url: https://your-domain.atlassian.net
It can also be used as part of an existing workflow by adding it as a step. More information about the options here.
action-jira-linter
adds a status check which helps you avoid merging PRs which
are missing a valid Jira Issue Key in the branch name. It will use the Jira
API to validate a given key.
When a PR passes the above check, action-jira-linter
will also add the issue
details to the top of the PR description. It will pick details such as the Issue
summary, type, estimation points, status and labels and add them to the PR
description.
action-jira-linter
will automatically label PRs with:
- A label based on the Jira Project name (the project the issue belongs to). For
example, if your project name is
Escher
then it will addescher
as a label. HOTFIX-PROD
- if the PR is raised againstproduction-release
.HOTFIX-PRE-PROD
- if the PR is raised againstrelease/v*
.- Jira issue type (based on your project).
Issue status is shown in the Description.
Why validate issue status?
In some cases, one may be pushing changes for a story that is set to
Done
/Completed
or it may not have been pulled into working backlog or
current sprint.
This option allows discouraging pushing to branches for stories that are set to
statuses other than the ones allowed in the project; for example - you may want
to only allow PRs for stories that are in To Do
/Planning
/In Progress
states.
The following flags can be used to validate issue status:
validate-issue-status
- If set to
true
,action-jira-linter
will validate the issue status based onallowed-issue-statuses
- If set to
allowed-issue-statuses
- This will only be used when
validate-issue-status
istrue
. This should be a comma separated list of statuses. If the detected issue's status is not in one of theallowed-issue-statuses
thenaction-jira-linter
will fail the status check.
- This will only be used when
Example of invalid status
<p>:broken_heart: The detected issue is not in one of the allowed statuses :broken_heart: </p>
<table>
<tr>
<th>Detected Status</th>
<td>${issueStatus}</td>
<td>:x:</td>
</tr>
<tr>
<th>Allowed Statuses</th>
<td>${allowedStatuses}</td>
<td>:heavy_check_mark:</td>
</tr>
</table>
<p>Please ensure your jira story is in one of the allowed statuses</p>
action-jira-linter
will add comments to a PR to encourage better PR practices:
A good PR title
When the title of the PR matches the summary/title of the issue well.When the title of the PR is slightly different compared to the summary/title of the issue
When the title of the PR is very different compared to the summary/title of the issue
A comment discouraging PRs which are too large (based on number of lines of code changed).
Batman says no large PRs 🦇A full example with all available options and example values is provided below.
- uses: btwrk/[email protected]
with:
github-token: ${{ secrets.GITHUB_TOKEN }}
jira-token: ${{ secrets.JIRA_TOKEN }}
jira-base-url: https://your-domain.atlassian.net
skip-branches: '^(production-release|master|release\/v\d+)$'
skip-comments: true
pr-threshold: 1000
validate-issue-status: true
allowed-issue-statuses: |
To Do
In Progress
Done
fail-on-error: false
Key | Description | Required | Default |
---|---|---|---|
github-token |
Token used to update PR description. GITHUB_TOKEN is already available when you use GitHub actions, so all that is required is to pass it as a param here. |
x | null |
jira-token |
Token used to fetch Jira Issue information. Check below for more details on how to generate the token. | x | null |
jira-base-url |
The subdomain of JIRA cloud that you use to access it. Ex: https://your-domain.atlassian.net . |
x | null |
skip-branches |
A regex to ignore running action-jira-linter on certain branches, like production etc. |
'' |
|
skip-comments |
A Boolean if set to true then action-jira-linter will skip adding lint comments for PR title. |
false |
|
pr-threshold |
An Integer based on which action-jira-linter will add a comment discouraging huge PRs. |
800 |
|
validate-issue-status |
A Boolean based on which action-jira-linter will validate the status of the detected jira issue |
false |
|
allowed-issue-statuses |
A line-separated list of acceptable Jira issue statuses. The detected jira issue's status will be compared against this list and if a match is not found then the status check will fail. Note: Requires validate-issue-status to be set to true . |
'In Progress' |
|
fail-on-error |
A Boolean which, if set to true , fails the GitHub Action when an error occurs. Default true . |
false |
Special note on jira-token
: Since tokens are private, we suggest adding
them as GitHub secrets.
The Jira token is used to fetch issue information via the Jira REST API. To get the token you need to:
- Generate an API token via JIRA.
- Add the generated token as the secret
JIRA_TOKEN
in your GitHub project.
Note: The user needs to have the required permissions (mentioned under GET Issue).
Since GitHub actions take string inputs, skip-branches
must be a regex which
will work for all sets of branches you want to ignore. This is useful for
merging protected/default branches into other branches. Check out some examples
in the tests.
action-jira-linter
already skips PRs which are filed by bots (for eg.
dependabot). You can add more bots to this list, or add the
branch-format followed by the bot PRs to the skip-branches
option.
Follow the instructions here to know more about GitHub actions.
Before committing a change - ensure that npm run all
has been executed and the
built js files are pushed to the repository as well.
Why is a Jira key required in the branch names?
The key is required in order to:
- Automate change-logs and release notes ⚙️.
- Automate alerts to QA/Product teams and other external stake-holders 🔊.
- Help us retrospect the sprint progress 📈.
Is there a way to get around this?
Nope 🙅Thanks goes to these wonderful people (emoji key):
Raj Anand 💻 👀 🤔 |
Aditi Mohanty 💻 📖 🚇 |
Dustin Row 👀 |
richardlhao 💻 |
Nimesh Manmohanlal 📖 |
Lewis Waddicor 💻 |
Asbjørn Ulsberg 💻 |
This project follows the all-contributors specification. Contributions of any kind welcome!