/cloudify-action-rest

Primary LanguageShell

cloudify-action-rest

This blueprint example shows how to run a GitHub Actions workflow from Cloudify and wait until it completes. The example blueprint runs the workflows in the .github/workflows directory, which are all just simple scripts, one after the other:

  • run-basic-script pipeline: Runs a basic script that has no delay
  • run-script-with-delay: Runs a script with an artificial delay to demonstrate how the logic works to wait for completion
  • run-script-with-inputs: Runs a single workflow command that accepts an input.

Explanation

The "wait until workflow is completed" logic is complicated by the nature of GitHub's API. Creating a workflow dispatch event is asynchronous, and does not return any information, other than an HTTP response. Therefore, the logic to run a pipeline and wait for it to finish must actually execute 4 API calls:

  1. Get the ID of the previous workflow run (see here)
  2. Initiate a new workflow run (see here or here with inputs)
  3. Obtain the ID of the newly executed workflow and make sure that it is not the same as the previous run (see here)
  • This avoids the race condition where a call to the list API does not yet contain the ID of the newly executed workflow from step 2 (again, this is a limitation of GitHub's API)
  1. Check the status of that workflow, and wait for it to complete (see here)

Each of these API calls is implemented as its own lifecycle operation in the sample blueprint. Each API call requires its own lifecycle operation for two reasons:

  1. Inputs must be passed between the steps (e.g., the previous run ID must be passed from step 1 to step 3)
  2. Placing all API calls into the same operation would retry all of the calls on failure of one of the calls. For example: if the call to check status failed, the API call to execute the workflow would also be retried

Usage

  1. Create a github_token secret in your Cloudify manager with repo scope permissions
  2. Upload and execute the blueprint, changing the inputs to match your repository location.

Caveat

The workflow that you are attempting to run must have been run at least once. Otherwise, the call to retrieve a previous workflow ID (step 1) will fail with an empty list.