Skip to content

build.md

Latest commit

 

History

History
48 lines (35 loc) · 1.54 KB

build.md

File metadata and controls

48 lines (35 loc) · 1.54 KB

Build and Push to ECR Workflow

This GitHub Actions workflow automates the process of building a Docker image from your BentoML service and pushing it to Amazon Elastic Container Registry (ECR).

Workflow Overview

This workflow performs the following steps:

  • Checkout the repository code.
  • Configure AWS credentials for deployment.
  • Login to Amazon ECR.
  • Install Python.
  • Set a short SHA of the commit in the environment.
  • Install required libraries.
  • Download models (optional).
  • Build the BentoML service.
  • Clean package cache.
  • Run Docker system prune to free up space.
  • Containerize the BentoML service.
  • Tag and push the Docker image to Amazon ECR.

How to Run

This workflow is triggered manually using the workflow_dispatch event. To run the workflow:

  1. Go to the Actions tab in your GitHub repository.
  2. Select the Build and Push to ECR workflow.
  3. Click on the Run workflow button

The python version is set to 3.8 by default. In case you need to use any other version, you need to update in the .github/workflows/build.yaml file at this step

- name: Install python
  uses: actions/setup-python@v5
  with:
    python-version: 3.12.3

Variables and Secrets

Required Variables

These variables need to be defined in the repository settings under Settings > Secrets and variables > actions:

  • ECR_REGISTRY_URI: The URI of your ECR repository (e.g.,XXXXXXXXXXXX.dkr.ecr.ap-south-1.amazonaws.com/iris-classifier-service). The iris_classifier_service will needs to be replacedvwith your service name.