Skip to content

Latest commit

 

History

History
55 lines (39 loc) · 3.35 KB

README.md

File metadata and controls

55 lines (39 loc) · 3.35 KB
Raw

Prefect Build Workflow

This GitHub Actions workflow automates the deployment process for Python applications using Prefect Cloud.
The workflow is designed to handle different environments (Production and Development) based on the branch where the changes are pushed.

Reviewing .github/workflows, there are two different possible GitHub actions, which match different triggers.
This is intended to provide different Prefect deployment mechanisms (.deploy() and .build_from_flow()).

Workflow Breakdown

Environment Variables

The workflow uses several environment variables, including PREFECT_API_KEY, PREFECT_API_URL, PROD_WORKSPACE, and DEV_WORKSPACE.
These are set as secrets in the repository settings for security.
With the API key and URL set, a prefect cloud login is not necessary.
Further, differentiating between workspaces (dev and prod or however your account is established) is only necessary if multiple workspaces are being utilized.
Your workspaces can be determined via prefect cloud workspace ls and should be saved in the form like chrisbprefectio/dev.

Trigger

There are two separate GitHub action workflows in this repository:

  1. prefect-deploy-prod - intended to trigger on pushes to main, and uses .build_from_flow to deploy to Prefect.
  2. prefect-deploy-feature - intended to trigger on pushes to feature-*, and uses .deploy to deploy to Prefect.

Jobs

Both workflow contains a single job named deploy, which runs on the latest Ubuntu runner.

Steps:

  1. Check out repository: The workflow checks out the code repository to make it accessible in the job.

  2. Set up Python 3.10: Python 3.10 is set up for the job, along with pip caching, to speed up dependency installation.

  3. Install Dependencies: Installs the required Python dependencies from requirements.txt.

  4. Environment-specific Actions:

    • Production Environment (Main branch):

      • Switches to the production workspace using the PROD_WORKSPACE environment variable.
      • Executes deployment.py to deploy on every push to main.

    • Development Environment (Feature branches):

      • Switches to the development workspace using the DEV_WORKSPACE environment variable.
      • Checks for and runs all modified Python files in feature branches, to deploy to resolved branches.

Environment Context

The BRANCH_REF environment variable is set dynamically to hold the reference of the branch where the workflow is running. The WORK_POOL variable is added to environment variables to resolve which worker should retrieve flow runs at run time. dev and prod are used in this example, but these should be changed to reflect the work pools used in your environment.

Usage

Push your changes to the appropriate branch (main for production, feature-* for development). The workflow will automatically handle the deployment process based on the branch context.

Notes

  • Ensure that the required secrets are set in the GitHub repository settings.
  • The workflow is designed to be flexible and can be adapted to different Python-based deployment strategies with Prefect Cloud.