Last updated Nov 21, 2022

Set up continuous delivery for Forge apps

This tutorial describes how to create a continuous delivery workflow for your Forge app. This will include a reference Bitbucket Cloud pipeline configuration for a hello world app built in Forge.

We’ll demonstrate how to:

  • Use environment variables to authenticate Forge CLI commands.
  • Orchestrate deployment processes triggered by Git merges.
  • Create deployment workflows for staging and production environments.

This tutorial is suitable for readers already familiar with CI/CD concepts and practices. You’ll learn useful tips specific to Forge app development. If you’re new to CI/CD, we recommend that you start with Continuous integration vs. delivery vs. deployment.

Before you begin

To complete this tutorial, you’ll first need to:

  • Build a Forge app. We built a reference app using our Confluence hello world app tutorial.
  • Install the app to an Atlassian site where you have admin rights.
  • Import the app’s code to a Git repository that features CI/CD workflows. In this tutorial, we use Bitbucket Cloud, and the application files are in the repository’s root folder.

You need to keep your app’s login details handy. This is the Atlassian API token and login email of your app’s owner.

While this tutorial is largely focused on Bitbucket Cloud pipelines, we also provide general guidance for GitHub users.

Step 1: Create a pipeline from scratch

  1. Create a new file named bitbucket-pipelines.yml.

  2. Forge apps are based on Node.js, so you need to define a Node docker image for your workflow. Set your pipeline’s Node.js version by adding the following line:

    1
    2
    image:
      name: node:18
    
  3. Add a code linting step to check your app’s code for common errors:

    1
    2
    pipelines:
      default:
        - step:
            name: Check Code linting
            script:
              - npm install
              - npm install --global @forge/cli
              - forge settings set usage-analytics true
              - forge lint
            caches:
              - node
    
  4. Save the file and add it to your repository’s root directory.

With a pipeline configuration added to your repository, you can now enable pipelines. Go to Repository settings > Pipelines > Settings to do so:

Enable Pipelines for your repository

For more detailed information about pipelines, refer to the Bitbucket Cloud documentation.

Understanding usage analytics

Every step in this tutorial’s reference pipeline includes the following command:

1
2
forge settings set usage-analytics true

This command provides the consent required by Forge to collect data about your app's deployments and installations (including error data). This, in turn, helps us monitor the overall performance and reliability of Forge. The collected data also helps us make better decisions on improving Forge's feature set and performance.

For information about how Atlassian collects and handles your data, read our Privacy Policy.

If you prefer not to enable usage analytics, change the setting to false.

Step 2: Define your login details as variables

Your Bitbucket pipeline will deploy and install your Forge app; both steps require your login details. The Forge CLI lets you store your login email and Atlassian API token through the environment variables FORGE_EMAIL and FORGE_API_TOKEN.

In CI/CD environments, you can store these variables as secrets for your builds. In your Bitbucket Cloud repository, you can define these variables in Repository settings > Pipelines > Repository variables.

Define your Forge login details through repository variables

When you define FORGE_API_TOKEN, make sure to tick Secured. This will ensure that your token won’t show up on build logs.

For detailed information about deployment variables in Bitbucket Cloud, refer to Variables and secrets.

Step 3: Configure continuous delivery workflow

After defining your login details as repository variables, you can define your deployment steps. These steps will use forge deploy and forge install commands, both of which require your login details.

Deploy to Staging

Your Bitbucket pipeline now has one step (forge lint). You can add another step to deploy and install the app to your staging environment:

1
2
        - step:
            name: Deploy to Staging
            deployment: staging
            caches:
              - node
            script:
              - npm install
              - npm install --global @forge/cli
              - forge settings set usage-analytics true
              - forge deploy -e staging
              - forge install --upgrade --site MYSITE.atlassian.com --product confluence --non-interactive -e staging

With this step, note the following:

  • Replace MYSITE.atlassian.com with your Atlassian site’s domain name.
  • In the previous step, you securely defined your login details through the FORGE_EMAIL and FORGE_API_TOKEN environment variables. The Forge CLI will use these automatically, so you don’t need to run forge login.
  • You use forge install --upgrade since the app is already installed on the site. Refer to the Forge CLI reference for more information about the options used in forge install.

Deploy to Production

Add a new step for deploying the app to your production environment:

1
2
        - step:
            name: Deploy to Production
            trigger: manual
            deployment: production
            caches:
              - node
            script:
              - cd hello-world-app
              - npm install
              - npm install --global @forge/cli
              - forge settings set usage-analytics true
              - forge deploy -e production

After you’ve added both pipeline configurations, commit the change. This will immediately run your pipeline. To view its progress through your Bitbucket Cloud repository, go to Pipelines:

View pipeline progress

Once the pipeline Status is Successful, select it to view its details. Notice that the Deploy to Production step didn’t run:

Manually deploy to production

This is because the Deploy to Production step uses the trigger: manual option. To manually run the step, click Deploy. Manual triggers help your team control changes to production while also automating all deployments to staging.

Reference pipeline

Refer to the following sample for the complete contents of this tutorial’s bitbucket-pipelines.yml file:

1
2
image:
  name: node:18

pipelines:
  default:
    - step:
        name: Check Code linting
        script:
          - npm install
          - npm install --global @forge/cli
          - forge settings set usage-analytics true
          - forge lint
        caches:
          - node
              
    - step:
        name: Deploy to Staging
        deployment: staging
        caches:
          - node
        script:
          - npm install
          - npm install --global @forge/cli
          - forge settings set usage-analytics true
          - forge deploy -e staging
          - forge install --upgrade --site MYSITE.atlassian.com --product confluence --non-interactive -e staging
          # remove this forge install step if you prefer to deploy 
          # major app changes through your site's Universal Plugin Manager.

    - step:
        name: Deploy to Production
        trigger: manual
        deployment: production
        caches:
          - node
        script:
          - npm install
          - npm install --global @forge/cli
          - forge settings set usage-analytics true
          - forge deploy -e production

Your pipelines should run npm and forge commands from your application’s root folder. This is the folder where your manifest.yml and package.json files are located. For example, if both files are in the repository's hello-world-app folder, add the following command before npm install in each step:

1
2
- cd hello-world-app

Rate this page: