Skip to content

Latest commit

 

History

History
134 lines (99 loc) · 11.2 KB

README.md

File metadata and controls

134 lines (99 loc) · 11.2 KB

CI/CD Pipeline

Technical level: Complex
Industry: Any
Use Cases: Modeler Building Blocks

Summary

Camunda can also be used to orchestrate your process development life cycle (PDLC), covering all steps from change reviews to release activities. This blueprint demonstrates a customizable CI/CD pipeline for deploying the contents of a Web Modeler folder to different environments using GitLab and Camunda 8. To learn more about the topic, review Integrate Web Modeler into CI/CD. During the process, a merge request is created in the target repository, which, when merged, triggers the deployment to the development environment. For deploying to the stage and production environments, custom pipeline runs have to be triggered. The blueprint is designed to be used with GitLab, but can be adapted to work with other CI/CD tools.

How to use

Set up Camunda 8 cluster

SaaS

  1. Create an API client as described in the documentation (with at least "Get" and "Create" scopes for "Web Modeler API").
  2. Camunda can also be used to orchestrate your process development lifecycle (PDLC), covering all steps from change reviews to release activities.. Click on "Create" and create a connector secret WEB_MODELER_CLIENT_ID and WEB_MODELER_CLIENT_SECRET in the cluster you want to use the blueprint in and set the values to the "Client ID" and "Client Secret" respectively.

Self-Managed

  1. Create an M2M application for the Web Modeler API as described in the documentation with at least create:* and read:* permissions.
  2. Follow the documentation to create the connector secrets WEB_MODELER_CLIENT_ID and WEB_MODELER_CLIENT_SECRET in the cluster you want to use the blueprint in and set the values to the "Client ID" and "Client Secret" respectively.

Set up the CI/CD blueprint in Web Modeler

  1. Import the blueprint diagram into a Web Modeler project.
  2. Navigate to the start event and open check the "Public access enabled" box in the properties panel in the "Publication" section.
  3. Deploy the diagram to the target Camunda 8 cluster.

Configure the target GitLab repository

Set up webhook

  1. Within the blueprint's diagram, navigate to the "Check pull request status" task, open the "Webhook" tab in the properties panel, and copy the Webhook path.
  2. Create a new GitLab repository/navigate to the repository you want to use as target for your CI/CD pipeline.
  3. In the repository settings, navigate to the Webhooks section and click on "Add new webhook".
    1. Set URL to the one you copied in step 1.
    2. Assign a secret token to the webhook.
    3. In the section "Trigger", select "Merge request events".

Repeat the steps for all Webhook Intermediate Catch Event Connectors within the "Deploy file" subprocess.

Important

By default, the same secret is used for all connectors. If you want to use different secrets, you have to adjust the blueprint accordingly.

After setting up the GitLab repository, create a new connector secret GITLAB_WEBHOOK_SECRET in the cluster you want to use the blueprint in and set the value to the secret you assigned to the GitLab webhook.

Set up deployment pipeline

  1. Copy the pipeline from the gitlab-pipeline directory to your target repository root.
  2. Within the repository settings, navigate to "CI/CD > Variables" and add the following variables to the target repository:
Development Environment
Variable name Description
DEV_ZEEBE_ADDRESS The address of the production Zeebe cluster, e.g. 3edeb612-7423-4683-b85b-095f9b90ab64.bru-2.zeebe.camunda.io:443
DEV_ZEEBE_CLIENT_ID The ID of the Zeebe client used for deploying the diagram to the development environment.
DEV_ZEEBE_CLIENT_SECRET The ID of the Zeebe client used for deploying the diagram to the development environment.
DEV_WEB_MODELER_WEBHOOK_URL The URL of the webhook reporting back to the blueprint.
DEV_WEB_MODELER_WEBHOOK_SECRET The secret you want to use for the webhook reporting back to the blueprint. Defaults to use the same secret as configured above.
Stage Environment
Variable name Description
STAGE_ZEEBE_ADDRESS The address of the production Zeebe cluster, e.g. 3edeb612-7423-4683-b85b-095f9b90ab64.bru-2.zeebe.camunda.io:443
STAGE_ZEEBE_CLIENT_ID The ID of the Zeebe client used for deploying the diagram to the staging environment.
STAGE_ZEEBE_CLIENT_SECRET The ID of the Zeebe client used for deploying the diagram to the staging environment.
STAGE_WEB_MODELER_WEBHOOK_URL The URL of the webhook reporting back to the blueprint.
STAGE_WEB_MODELER_WEBHOOK_SECRET The secret you want to use for the webhook reporting back to the blueprint. Defaults to use the same secret as configured above.
Production Environment
Variable name Description
PROD_ZEEBE_ADDRESS The address of the production Zeebe cluster, e.g. 3edeb612-7423-4683-b85b-095f9b90ab64.bru-2.zeebe.camunda.io:443
PROD_ZEEBE_CLIENT_ID The ID of the Zeebe client used for deploying the diagram to the production environment.
PROD_ZEEBE_CLIENT_SECRET The ID of the Zeebe client used for deploying the diagram to the production environment.
PROD_WEB_MODELER_WEBHOOK_URL The URL of the webhook reporting back to the blueprint.
PROD_WEB_MODELER_WEBHOOK_SECRET The secret you want to use for the webhook reporting back to the blueprint. Defaults to use the same secret as configured above.

Generate an access token

  1. Navigate to "Settings > Access Tokens" and create a new token for the repository you want to use the blueprint in:
    • Scopes: api, write_repository
    • Roles: Developer
  2. Save the token as a new connector secret GITLAB_PAT within your cluster.

Run the blueprint

You can either start the blueprint via Camunda Tasklist or via running it directly in Web Modeler. Additionally, on Camunda 8 SaaS, you can use the public link of the blueprint's start event to start a new instance of the blueprint.

The form will ask you for the following information:

  • Folder ID: The ID of the folder you want to deploy to the target repository. You can find the ID in the URL of the folder in Web Modeler: Folder ID
  • Repository URL: The URL of the target repository.
  • Target Base Branch: In the course of the blueprint, a new branch will be created in the target repository. This field specifies the base branch for the new branch.
  • Target GitLab Base URL: The base URL of the GitLab instance. Defaults to https://gitlab.com.
  • Target GitLab Project ID: The ID of the target project in GitLab. You can find the ID in the URL of the project in GitLab: Project ID

After starting a process instance, it will create a pull request with the contents of the target folder in the target repository. When you merge the pull request, the folder contents will be deployed to the development environment via the deploy_to_dev job within the pipeline. To check the status of the process, you can always use Camunda Operate to track the status of the process. Operate should indicate with blue colored sequence flows where the process went and mark the current activity with a green circle:

Operate Process Instance

For deploying to the stage environment, create a custom pipeline run via "Build > Pipelines > Run pipeline" and additionally pass a variable ENVIRONMENT with the value stage to the pipeline run:

Deploy to staging environment

For deploying to the production environment, pass a variable ENVIRONMENT with the value prod to the pipeline run.

Known Limitations

  • Concurrent runs of the blueprint are not supported. This means that if you start the blueprint while another instance is still running, deployment information might be mixed between the two instances.
  • Changes to the created branch are not tracked. This means that if you change the content of a synced file within the git repository, the blueprint will not update the file content within Web Modeler.
  • Adapting the blueprint to work with GitHub is currently not supported. The Connectors Framework at the moment doesn't support Base64 encoding for process variables which is required by GitHub's "Create or update file contents" API.