Putting the Infrastructure to the Test

With all the infrastructure components now in place, it’s time to perform an end-to-end test run.

Test Plan

The following outlines a basic test plan:

• Create a GitHub repository within the organization where the GitHub to EventBridge integration has already been implemented
• Within the same repository, create a workflow definition, consisting of two (2) jobs
• Ideally, both jobs should be designed to execute in parallel
• Arbitrary runner label(s) can be assigned to each job
• Run the workflow

Expected Result(s)

After initiating a workflow run, we should expect to observe the following for each job:

• Actions workflow run summary displays status of “queued“
• GitHub webhook forwards event payload to the target AWS event bus
• EventBridge rule triggers the ECS Fargate JIT organization runner (container)
• Job executes within the ECS runner container
• Upon job completion, the runner automatically removes itself from the associated GitHub organization
• Runner container automatically stops/exits

Test Execution

The following resource/object names will be referenced during the testing procedure.

Test Objects

The workflow file listed in the previous table was created with the following contents:

name: AWS Fargate JIT test

on:
  push:

jobs:
  run-it:
    strategy:
      matrix:
        job_num: [1, 2]
    runs-on:
      - self-hosted
      - ${{ matrix.job_num }}
    steps:
      - name: Echo job_num
        run: |
          echo executing job_num: ${{ matrix.job_num }}

Trigger Workflow

Trigger a run by committing/pushing the workflow file to the remote repository.

Validate Test Run

• The actions run summary shows two (2) concurrent jobs with a status of “queued“

• Behind the scenes, the GitHub to EventBridge integration triggers the rule/targets required to instantiate the runner(s)
• Jobs wait for runner(s) to come online

• After checking the ECS Fargate cluster tasks list, we observe two tasks with status “provisioning“, which confirms that the EventBridge rule detected the events, and invoked the associated rule target(s)

• After a short while, the Fargate tasks transition from a state of “provisioning” to “running“, and a message appears in the actions log to indicate the “Job is about to start running on the runner….“

• If all goes well, workflow run logs should show both jobs as successfully completed

Check the Fargate task container logs for each job to confirm the following:

• job ran successfully within the container
• the JIT runner automatically removed itself
• the container exited/stopped

Below is a sample Fargate task log for a successful job run: