CI/CD Pipeline using GCP Cloud Build

[jimmyzhen]

Note: This repository has existing CI integration with CircleCI at https://app.circleci.com/pipelines/github/MoTrPAC/exercise-transcriptome-meta-analysis

Primary objective:
Establish a continuous delivery pipeline that automates the portal deployment process as much as possible.

Approaches:

1. Leverage the existing (free tier) CircleCI integration to run builds and unit tests on any given branches. The intention is to prevent any failed builds from being deployed to a GKE cluster. Although, technically, the unit testing step can be included in the Cloud Build pipeline, I prefer it to be done ahead of time both locally and on CircleCI while letting Cloud Build to focus on deployment.
2. When a pull request of a working branch is merged into the dev branch, it triggers the Cloud Build to run the pipeline and deploy code changes to a GKE cluster specifically exposed for testing and demo purposes.
3. When a pull request of a working branch is merged into the main branch, it triggers the Cloud Build to run the pipeline and deploy code changes to a GKE cluster specifically exposed for staging and QA/QC purposes.
4. When the main branch is tagged with a release label, it triggers the Cloud Build to run the pipeline and deploy code changes to a GKE cluster specifically exposed for the production environment.

Considerations:

1. SSL certificates for GKE clusters are needed.
2. Testing and staging GKE cluster endpoints should potentially be inaccessible to the public.

[jimmyzhen]

Prior to opening this ticket, this repository had no other branches except the main branch.

I created the dev branch based off of the main branch and added a few config files to experiment with Cloud Build pipeline setup. The experiment went successfully and now I need to formally orchestrate the workflow and polish the configurations.