Configure consumer and provider pipelines

Setup deployment environment

Important

This step should only be required if you have a legacy API Hub for Contract Testing account. New users should automatically have a production account created and can move onto the Configure consumer pipeline step

Create a new production environment to record deployments against

  1. Log in to your API Hub for Contract Testing account (https://<your-subdomai>.pactflow.io), and go to Settings > Environments.

  2. Click Add Environment

  3. Enter production for the name and display name

  4. Check the "this is a production environment" checkbox

  5. Select the default team

  6. Click "Create"

Configure consumer pipeline

The source repositories are configured to use the the public broker at test.pactflow.io. You will need to update the credentials to point to your own API Hub for Contract Testing account. To do this, we need to update the PACT_BROKER_BASE_URL environment variable in the Github workflow file, and create a Github Secret to store the API Hub for Contract Testing API token in.

  1. Create a Github Secret to store your API Hub for Contract Testing API token in.

    • In API Hub for Contract Testing:

      • Log in to your API Hub for Contract Testing account (https://<your-subdomain>.pactflow.io), and go to Settings > API Tokens.

      • Click the Copy button for the read/write CI token (make sure it's the read write one, not the read only one).

    • In Github:

      • Open your forked example-consumer-legacy project (https://github.com/<your-username>/example-consumer-legacy)

      • Click on the Settings tab.

      • Select Secrets from the side menu.

      • Click New repository secret (the button is to the right of the "Actions secrets" heading)

      • Set the name of the secret to PACTFLOW_TOKEN_FOR_CI_CD_WORKSHOP

      • Paste in the API Hub for Contract Testing API token value you copied in the previous step.

  2. Update your workflow files in GitHub to point at your API Hub for Contract Testing Broker

    • In API Hub for Contract Testing

      • Go to Settings > API Tokens.

      • Click the COPY API Hub for Contract Testing BASE URL button

    • In Github:

      • Open your forked example-consumer-legacy project (https://github.com/<your-username>/example-consumer-legacy)

        • Open .github/workflows/build.yml

        • In the upper right corner of the file view, click 🖊️ to open the file editor.

        • Update the value of PACT_BROKER_BASE_URL to the base URL of your own API Hub for Contract Testing account. You can easily get this by clicking the COPY PACTFLOW BASE URL button on the API Tokens page in PactFlow.

        • Press the green Commit changes button

  3. View the build:

    • In Github:

      • Select the most recent build, this will have been triggered when you committed the changes in the last page

Important

This build should now successfully publish the pact, but it will fail on the can-i-deploy step before it tries to deploy.

This is because the provider has not published a successful verification result for the pact.

Configure provider pipeline

🔁 Repeat the above instructions to configure the API Hub for Contract Testing account for your provider project.

⚠️ There are TWO workflow files to be updated in the provider project - .github/workflows/build.yml and .github/workflows/verify_changed_pact.yml.

After you have pushed your changes to the workflow files, the provider pipeline will run, fetching and verifying the configured pacts from your API Hub for Contract Testing account, and publishing the results back. The can-i-deploy command will pass, and allow the provider to be deployed. ✅

Back to the consumer

✅ To make both your builds go green, we are going to retry can-i-deploy, now that the consumer is deployed to production.

  1. Find the latest failing example-consumer-legacy workflow in the Github Actions page (Actions -> Under Workflows, select Build -> failing build).

  2. In the top right of the failing job page, select to Re-run the failed jobs (can-i-deploy) (Actions -> Under Re-run jobs, select Re-run failed jobs).

  3. The test run should be pre-passing

  4. The can-i-deploy step should now pass - our consumer is safe to deploy, now that our provider published a verification result, and is deployed to production

    • If you click on the step you should see a successful verification result, and Computer says yes \o/

  5. The deploy step should pass too - our consumer is deployed to production, and we record this with the record-deployment command, marking this application version of our consumer as deployed to production

Expected state by the end of this step

  • Both consumer and provider builds passing ✅

In this section:
© 2024
Publication date: