This repository is a template and demo for DevOps/MLOps on AWS using GitHub Actions and Terraform. It includes workflows for CI/CD, infrastructure provisioning, and deployment of applications to Kubernetes clusters. Additionally the infrastructure includes a Prometheus and Grafana stack for monitoring.
Initiate the terraform.yaml GitHub actions workflow to provision AWS resources using Terraform. This workflow can be triggered manually from the GitHub website. This workflow will:
- Create an S3 bucket for storing Terraform state.
- Create a DynamoDB table for state locking.
- Create an EKS cluster with a node group.
flowchart TD
A[GitHub Actions] --> B[Terraform]
B --> C[EKS Cluster]
A --> F[AWS CLI]
F --> D[S3 Bucket for State]
F --> E[DynamoDB Table for State Locking]
Github Actions is used to deploy applications to an EKS cluster. The workflow will:
- Build and push Docker image to this repository's container registry.
- Deploy the application to the EKS cluster using the provided Kubernetes manifests.
flowchart TD
A[GitHub Actions] --> B[Docker Build & Push]
A --> C[EKS Deployment]
B --> D[Container Registry]
C --> E[Kubernetes Manifests]
TF_STATE_BUCKET: Name of the S3 bucket for Terraform state.TF_STATE_TABLE: Name of the DynamoDB table for state locking.AWS_ACCESS_KEY_ID: AWS access key ID.AWS_SECRET_ACCESS_KEY: AWS secret access key.KUBECONFIG: Base64 encoded kubeconfig file for accessing the EKS cluster.
AWS_REGION: The AWS region where resources will be created (e.g.,us-east-1).EKS_CLUSTER_NAME: The name of the EKS cluster to be created.
Generate kubeconfig using the following command once the EKS cluster is created:
aws eks update-kubeconfig --name <EKS_CLUSTER_NAME> --region <AWS_REGION> --profile <AWS_PROFILE>kubectl get nodesAWS IAM permissions for Terraform backend described here: https://developer.hashicorp.com/terraform/language/backend/s3