Deploy application containers to Amazon EKS with AWS App2Container

Amazon Elastic Kubernetes Service (Amazon EKS) is a managed service that you can use to run Kubernetes on AWS. Amazon EKS streamlines the provisioning of highly available and secure clusters, and automates key maintenance tasks such as patching, node provisioning, and updates. Kubernetes is an open-source system for automating the deployment, scaling, and management of containerized applications. For more information about Amazon EKS, see What is Amazon EKS? in the Amazon EKS User Guide.

Prerequisites

To configure an Amazon EKS integration for your application container with App2Container, your application must meet the following criteria.

• Java applications running on Linux must satisfy Java application framework requirements, and run on a supported Linux distribution. For details, see Supported applications, and expand the Supported applications for Linux section.

• .NET applications running on Windows must satisfy application framework and system requirements, and meet the criteria for supported applications. For details, see Supported applications, and expand the Supported applications for Windows section.

• Application containers that run in Amazon EKS must launch EC2 instances. App2Container does not currently support Fargate as a container launch type for Amazon EKS.

Amazon EKS integration for App2Container workflow

The process for setting up application containers for hosting in Amazon EKS is integrated smoothly with the App2Container workflow. Initial steps for App2Container are the same for all applications.

1. Install and set up the App2Container environment, as described in the Prerequisites: Set up your servers section.

2. Complete the initialization phase for your App2Container environment with the init command, and the remote configure command, if applicable. To learn more about what is included in all of the App2Container containerization phases, see the Command reference.

3. Complete the analyze phase for each application that you want to containerize.

   • If you are running commands directly on application servers, use the inventory and analyze commands.

   • If you are running a remote workflow on a worker machine, use the remote inventory and remote analyze commands.

4. Integration begins with the containerization step.

   • When you run the containerize command, App2Container generates the deployment.json file, which provides configurable parameters for all supported container management service options that could apply to your application container.

   • Parameters for Amazon ECS and Amazon EKS are always included. Parameters for App Runner are also included if your application container meets the App2Container criteria for hosting in App Runner

   • Each container management service has its own section in the deployment.json file, and each section has a flag to indicate which container management service is the destination for your application container. Only one section can have its flag set to true – all others must be set to false.

     Amazon ECS is configured as the destination by default. To deploy your application containers to Amazon EKS, you can set the createEksArtifacts in the eksParameters section to true, and the createEcsArtifacts flag in the ecsParameters section to false. For more information about configuring the deployment.json file, see Configure deployment.

   • App2Container configures HTTP-based deployments by default. To use HTTPS for your deployment, specify the Amazon Resource Name (ARN) of an AWS Certificate Manager (ACM) certificate in the deployment.json file. For more information, see Configure deployment.

5. The deployment step creates artifacts that are targeted for the Amazon EKS container hosting service, based on the settings in the deployment.json file, where the createEksArtifacts flag is set to true.

   • When you run the generate app-deployment command, App2Container validates the properties in the deployment.json file, and pushes the container image to Amazon ECR. This is the standard workflow.

   • The command generates a CloudFormation template (eks-master.yml) that creates an EKS cluster, pulls your application container images from Amazon ECR, and deploys your application to the cluster.

     It also generates Kubernetes manifests (eks_deployment.yaml, eks_service.yaml, and eks_ingress.yaml), for post-deployment customizations using a tool such as kubectl.

   • The command generates the pipeline.json file to support creating a pipeline to deploy updates to your application container in Amazon ECR.

   • If you use the --deploy option for the generate app-deployment command, App2Container deploys the AWS CloudFormation stack that creates the Amazon EKS service for the containerized application, using the configuration values in the AWS CloudFormation template that it generates. To customize the configuration, run the command without the --deploy option, and then manually deploy using the AWS CLI when you are ready.

6. The pipeline step generates a CloudFormation template for the pipeline that is targeted for the Amazon EKS container management service, based on the settings in the pipeline.json file.

   • When you run the generate pipeline command, App2Container validates the properties in the pipeline.json file, and verifies that initial deployment to Amazon EKS has been completed, and that your application is active.

   • The command generates a CloudFormation template to create a two-step pipeline:

     1. Code commit – Creates or updates an AWS CodeCommit repository that contains the Dockerfile and application artifacts that are required to create your application container image.

     2. Code build – Builds the Docker image for your application container, and pushes the updated image to the Amazon ECR repository that you configured for your application.

     3. If you use the --deploy option for the generate pipeline command, App2Container deploys the pipeline with the configuration values in the CloudFormation template it generates. To customize the configuration, run the command without the --deploy option, and then manually deploy using the AWS CLI when you are ready.