Deploy an ECS Fargate service

Steps

In this example, the source code in the directory Kattehotell.Api/ is deployed to your AWS account:

.
└── Kattehotell.Api/
    ├── ...
    └── Dockerfile

This ECS Fargate is deployed as part of the application Kattehotell.

Step 1: Prepare an ECR repository

1. Navigate to repository SharedKontoInfrastruktur
2. Navigate to ./Shared/<application-name>/eu-west-1/ecr/main.tf
3. Update locals with new entries:

   ./Shared/kattehotell/eu-west-1/ecr/main.tf

   1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23

   locals { kattehotell_repos = toset([ ... "api", ]) account_access = [ local.accounts["BYM-DP-Kattehotell-Test"].id, local.accounts["BYM-DP-Kattehotell-Prod"].id ] } module "kattehotell_ecr" { for_each = local.kattehotell_repos source = "git@github.com:BYM-IKT/terraform-aws-ecr.git?ref=v3" application_name = "kattehotell" service = each.value account_access = local.account_access } output "kattehotell_ecr_repo" { value = module.kattehotell_ecr }

4. Create a Pull Request to Terraform apply this change. Terraform provisions a new ECR repository with the name ${application_name}-${service}. In this example, the resulting ECR repository receives the name kattehotell-api.

Step 2: Build and push Docker Image

Before you can proceed, the ECR repository cannot be empty. To populate it with an image, navigate to your product's repository, and do the following steps:

1. Create a GitHub Actions workflow file .github/workflows/deploy-to-test.yml:

   .
   ├── .github/
   │   └── workflows/
   │       └── deploy-to-test.yml
   └── Kattehotell.Api/
       ├── ...
       └── Dockerfile
   

   with the following content:
   ./.github/workflows/deploy-to-test.yml

   name: Deploy ECS Fargate to TEST
   
   on:
     push:
       branches: [master]
     workflow_dispatch:
   
   jobs:
     build_push_image_to_shared:
       name: Build image and publish to ECR
       runs-on: ubuntu-latest
       permissions:
         id-token: write
         contents: read
       steps:
         - name: "Kattehotell.Api"
           uses: BYM-IKT/github-actions/build-and-push-image-to-ecr@master
           with:
             dockerfile-path: Kattehotell.Api/Dockerfile
             aws-account-id:  "<<AWS_ACCOUNT_ID>>"        
             ecr-name:        kattehotell-api
             image-tags:      latest
   
     deploy_images_to_ecs_service_test:
       needs: [build_push_image_to_shared]
       name: "Deploy to TEST"
       uses: BYM-IKT/github-actions/.github/workflows/deploy-image-to-ecs.yml@master
       with:
         environment:           testing
         aws-account-id-target: "<<AWS_ACCOUNT_ID>>"
         ecr-name:              kattehotell-api
         image-tag-target:      latest
         image-tag-new:         test
         ecs-cluster-name:      kattehotell
         ecs-service-name:      api
   

   where <<AWS_ACCOUNT_ID>> is replaced with the Id of your AWS Account.

2. Run the pipeline, which builds and pushes your Docker image to your ECR repository.

Step 3: Create an ECS Fargate service

In your product's Infrastraktur-repo, provision your ECS Fargate service referencing the ECR repository's URI as follows:

module "application" {
  source = "git@github.com:BYM-IKT/terraform-byks-module.git?ref=v12"
  environment = var.environment
  ...
  ecs_services = {
    api = {
      ...
      ecr_uri   = data.terraform_remote_state.shared_ecr.outputs["api"]
      image_tag = var.environment
    }
  }
}

Create a Pull Request to Terraform apply this change.

Terraform provisions a new ECS Fargate service named ${environment}-${application_name}-${service}. In this example, the resulting name of the ECS Fargate service is api and it resides in the kattehotell cluster.

Step 4: Retry application pipeline

The whole pipeline created in Step 2 should work now. Try rerunning it.