Marketplace homepage: rainforestapp/github-action
This is the Rainforest QA GitHub Action, it allows you to easily kick off a Rainforest run from your GitHub workflows, to make sure that every release passes your Rainforest integration tests.
Sections
Prerequisites
A Rainforest QA account
If you don't already have one, you can talk to us about setting up an account here.
A Rainforest QA API token
You can find yours on the Integrations setting page.
Do not expose this token in your .github/workflows/<workflow>.yml file. Instead, use a GitHub encrypted secret. You may name this secret as you wish (as long as it's a name allowed by GitHub), e.g. RAINFOREST_API_TOKEN.
A run group with at least one test
Run groups are a way to group tests that should be run together (for example, a smoke suite you might want to run on every deploy). For more information on run groups, see this help article.
Once you have a run group which contains at least one enabled test, you can run it in GitHub workflows using this Action. You will need its ID (visible at the end of the run group URL: https://app.rainforestqa.com/run_groups/<ID>).
Base usage
This is a simple workflow file that will start a Rainforest run using run group #1234 every time a commit is pushed to the main branch. To use it, change the run group ID to the one you wish to run, ensure your Rainforest API token is properly configured in your GitHub secrets, and commit this file in your repo at .github/workflows/rainforest.yml.
on: push: branches: - main jobs: rainforest: runs-on: ubuntu-latest name: Run Rainforest steps: - name: Rainforest uses: rainforestapp/github-action@master with: token: ${{ secrets.RAINFOREST_API_TOKEN }} run_group_id: 1234
Optional Parameters
description
An arbitrary string to associate with the run.
Default behavior
The default value :
"${GITHUB_REPOSITORY} - ${GITHUB_REF_NAME} ${GITHUB_JOB} $(date -u +'%FT%TZ')"
This means that if no description parameter is passed in and your repository is named my_org/my_repo, the GitHub job is #42 on the my_feature_branch branch, and the current time (in UTC) is noon on December 31st, 2021; then the created run's description will be:
my_org/my_repo - my_feature_branch 42 2021-12-31T12:00:00Z
environment_id
Use a specific environment for this run. This parameter will be ignored if the custom_url parameter is also passed in.
Default behavior
If no environment_id parameter is passed in, the created run will use the Run Group's default environment.
custom_url
Use a specific URL (via a temporary environment) for this run.
Default behavior
If no custom_url parameter is passed in, the created run will use the Run Group's default environment.
conflict
How we should handle currently active runs.
Allowed values
| Value | Behavior |
|---|---|
cancel |
Cancel all other runs in the same environment. |
cancel-all |
Cancel all other runs, regardless of environment. |
Default behavior
If no conflict parameter is passed in, then no active runs will be canceled.
execution_method
The execution method to use for this run.
Allowed values
| Value | Behavior | Requirements |
|---|---|---|
automation |
Run against our automation agent. | - All tests in the run group are written with the Visual Editor. - No tests use a Tester Instruction/Confirmation action. |
crowd |
Run against our global crowd of testers. | |
automation_and_crowd |
Run against our automation agent where possible, fall back to the crowd of testers. | |
on_premise |
Run against your internal testers. | - On-premise is enabled for your account. |
Default behavior
If no execution_parameter parameter is passed in, the created run will run against the run group's default execution method.
release
A string used to link a run to a release (for example, a git SHA or tag, a version number, a code name)
Default behavior
If no release parameter is passed in, the SHA1 hash of the latest commit of the current workflow (obtained via the GITHUB_SHA environment variable) will be used.
automation_max_retries
Set to a value larger than 0 to retry failed tests excuted by our automation agent in the same run up to that number of times.
Default behavior
If no automation_max_retries parameter is passed in, the default from your account or run group is used.
branch
Use a specific Rainforest branch for this run.
Default behavior
If no branch parameter is passed in, the main branch will be used.
background
Set to true to immediately complete the GitHub workflow job without waiting for the Rainforest run to complete
Default behavior
By default we wait for the run to complete in order to pass or fail the workflow job based on the run's result.
dry_run
Set to true to run parameter validations without actually starting a run in Rainforest.
Default behavior
If no dry_run parameter is passed in, the run will be started in Rainforest.
Rerunning failed tests
If your Rainforest run fails due to a "non-bug" (your testing environment might have had a hiccup, or a test might have needed to be tweaked, etc), then rather than make code changes and then run your full testing suite once more, you'll instead want to rerun just the tests that failed. The Rainforest QA GitHub Action uses GitHub caching to know when a workflow is being rerun. It will then automatically rerun only the tests which failed in the previous run.
Running multiple Rainforest runs from a single workflow using workflow_call
The Rainforest QA GitHub Action cannot detect on its own when it is being called multiple times in a single workflow through an intermediary reusable workflow. This means that the second invocation will attempt to rerun the run created by the first invocation, which will fail if that run has no failed tests to rerun. In order to avoid this situation, set the cache_key parameter to a distinct value in each invocation:
# reusable workflow on: workflow_call: inputs: cache_key: type: string jobs: rainforest: runs-on: ubuntu-latest name: Run Rainforest steps: - name: Rainforest uses: rainforestapp/github-action@master with: token: ${{ secrets.RAINFOREST_API_TOKEN }} run_group_id: 1234 cache_key: ${{ inputs.cache_key }}
# calling workflow jobs: first-invocation: uses: ./.github/workflows/reusable.yml with: cache_key: first-invocation second-invocation: uses: ./.github/workflows/reusable.yml with: cache_key: second-invocation
GitHub Action Release Process
This section describes the release process for the Action itself:
- Create a feature branch and do your work.
- Update the version in
action.yml. - Get review and merge to master.
- Create a GitHub Release with the proper
v-prefixed version tag (i.e.v0.0.1). List Bugfixes, Breaking changes, and New features (if present), with links to the PRs. See previous releases for an idea of the format we're using. - The
release.ymlworkflow will then run to update the major release tag. E.g. if your release was forv1.2.3, then it will automatically update thev1tag.
If you want to run an integration test, create a new branch in a repo of your choice and add a new workflow to.github/workflows/ with the action pointing to your commit.