Set Up Feature Branch Deployment

[kornel-filep]

Configure additional pipelines for feature branches. Feature branches match with the regex "[0-9]+-**" (Github Issue regex). This pipeline should be manual trigger only and should not be activated for the main branch.

Once a feature branch has been initially deployed, the URL established should remain constant. Along with this, we need to set up a separate pipeline for AWS CDK Stack cleanup, enabling us to tidy up the environment when necessary (destroy the whole stack).

Acceptance Criteria:

1. Github Actions pipeline is in place for feature branch deployments to AWS.
2. The pipeline should only trigger manually (workflow dispatch) for branches matching the "[0-9]+-**" regex.
3. It should not be available for the main branch.
4. The URL for a deployed feature branch remains constant after its first deployment.
5. Updates to a feature branch and manual deployment only change the website content, not the URL (after first deployment).
6. A second pipeline is available for complete AWS CDK Stack cleanup.

[kyouka5]

After pushing two pipelines to my branch, I have just noticed that I am not able to trigger them under the Actions tab. I did some digging, and found that pipelines with workflow_dispatch trigger can not be triggered as long as they are not in the main branch:

To trigger the workflow_dispatch event, your workflow must be in the default branch. For more information about configuring the workflow_dispatch event, see "Events that trigger workflows."

Source

A solution for this problem could be to change the trigger in the pipelines to push while we are still testing our changes, and then before opening a PR we could update the trigger to be workflow_dispatch.

@kornel-filep What do you think, maybe do you have any better ideas in mind?

FYI @t0mmili @taughty

[kornel-filep]

@kornel-filep What do you think, maybe do you have any better ideas in mind?

Sure that works for me @kyouka5.

[kyouka5]

After a discussion with @kornel-filep, it was decided that we are going to follow a different approach to feature branches than what we had initially considered. The changes are as follows:

• Instead of having multiple smaller stacks for each feature branch, we are going to have a single stack that manages all feature branches.
• Pipelines should be created for stack deployment and destroy.
• Feature branch deployment pipeline should use AWS CLI to store feature branches in separate directories in S3.
• Feature branch destroy pipeline should use AWS CLI to delete the feature branch S3 directory that was created in the deployment pipeline.

[kornel-filep]

@t0mmili please check the above message from @kyouka5 as it potentially has impact on the main branch pipelines.

The idea is that the main brach will be deployed on BASE_URL/ and the feature branches will be on BASE_URL/BRANCH/ endpoints. This potentially has implications on how the feature and main branch pipelines work in tandem.