05-docker-app-ecs

Container Games on Amazon ECS

• Containerized four web-based games using Docker with an Nginx-optimized base image, ensuring compatibility with AWS Fargate on the Linux x86_64 architecture.
• Pushed the Docker images to Amazon Elastic Container Registry (ECR) for secure and scalable image management.
• Deployed the containerized games on Amazon Elastic Container Service (ECS), ensuring seamless operation in a cloud-native environment.
• Implemented auto-scaling and load balancing to handle varying traffic loads efficiently and maintain high availability.
• Explored blue-green deployment strategies by manually deploying a new updated service, enhancing deployment practices for minimizing downtime and risks.
• Services : AWS ECS, ECR, ALB, Docker, Nginx

ECS GUIDE

Key Components of Amazon ECS

Introduction to Amazon ECS

• Amazon Elastic Container Service (ECS) is a fully managed container orchestration service provided by AWS that makes it easy to deploy, manage, and scale containerized applications.
• ECS supports Docker containers and allows you to run and manage containers on a cluster of EC2 instances or using AWS Fargate, a serverless compute engine.

1. Clusters:

• A logical grouping of container instances or tasks.
• Clusters can run on EC2 instances, AWS Fargate, or a mix of both.

2. Task Definitions:

• Blueprint for your application.
• Specifies parameters for the container, such as Docker image, CPU, memory, port mappings, and environment variables.
• Defines multiple containers within a single task if needed.

3. Tasks and Services:

• Tasks: A running instance of a task definition. It is the smallest unit of work in ECS.
• Services: Maintain the desired number of task instances. Can be associated with load balancers to distribute traffic across tasks.

4. Container Instances:

• EC2 instances registered to the ECS cluster.
• Run the ECS agent, which manages the tasks.

5. ECS Agent:

• A container running on each container instance that communicates with the ECS control plane to manage tasks.

6. Launch Types:

• EC2 Launch Type: Deploys tasks on a cluster of Amazon EC2 instances managed by the user.
• Fargate Launch Type: Deploys tasks on serverless infrastructure managed by AWS, removing the need to manage EC2 instances.

7. Task Scheduler:

• Places tasks based on the specified placement strategy and constraints.
• Ensures the desired state of tasks is maintained in the cluster.

8. Load Balancing:

• Integrates with Elastic Load Balancing (ELB) to distribute incoming traffic across tasks.
• Supports Application Load Balancer (ALB) and Network Load Balancer (NLB).

Docker Commands used in this mini project

$ docker build -t <image> .       # build
$ docker images                   # display build images
$ docker run -d -p 80:80 <image>  # run container
$ docker ps                       # display running containers
$ docker ps -a                    # display all/stopped  containers
$ docker stop <container-id>      # stop running container
$ docker rm <container-id>        # delete running container
$ docker rmi <image>              # delete image

---

Architecture

---

Output

Screenshots

Demo - Video : LINK

Cluster - Service

Cluster - Task

ALB

Blue Green Deployment

video demo
https://github.com/user-attachments/assets/97be68ad-b16a-4d88-a24c-88756a5e0216

changes

Force Deploy

Updating Cluster

Result Updated

Screenshots @ loacalhost:80 | Distroless Nginx

Docker Build

Docker Run

Result

---

Troubleshooting

1. Tried with Distroless Image : So ecs health check fails, due to absence of shell and curl in distroless nginx image Solution : Used Alpine Image

2. Architecture Problem : Mismatch Container of differenet architecture : amd64, arm64 can cause problem

   • Explanation :
     • My laptop's local builds resulted in arch:ARM containers (As I am Using MacBook Air M1).
     • And I selected Fargate as Arch: x86_64 ~ amd64. So containers were running and exiting.
     • So, have to troubleshoot in CloudWatch > Logs > exec failed
     • Solution: Build Multi-arch image or specify with --from arch explicitly
     • NOTE: On Fargate arm64, my image works, but i have used x86_64 in ECS Cluster setup.

WHY used x86_64 or amd64 ???
I am using Region : us-east !! for safe side