leejunkim

ECS,EKS

3 min read

ECS

  • Elastic Container/Kubernetes Services
    • ECS is AWS’s simpler container orchestrator, letting you define tasks and services without the complexity of Kubernetes.
    • EKS is the managed K8s solution, typically more advanced
  • All about running containers in the cloud
  • diagram
  • Services for running containerized workloads on AWS → u want to run images
  • Managed container orchestration with many configuration options
  • Handles cluster management, container deployment, scaling, and health monitoring
  • Alternative to EC2 (not recommended)
  • Perfect to fun any kind of containerized workloads

Deeper look

  • diagram
  • Define cluster structure
    • You have to define your cluster structure = the infrastructure environment where containers run
    • Which kind of infrastructure?
      • you could use EC2 instances or Fargate as container host
      • EC2 instances → these instances will be started and operated by ECS/EKS (and you won’t manage them manually). But you still have to set instance settings when you create a EC2/EKS
      • Fargate → recommended
    • Tasks: which image you should use & other image configs
    • Network & security settings
  • Operate & scale containers
    • you can define service/task specific settings
      • Need to select a Task definitions u want to use first
      • service - ongoing task (ex. web servers)
      • task - tasks finish at some time
    • monitor containers
    • stop/start
    • schedule tasks (ex. once everyday), or have something running all day (ex. web server)

Task definitions / Task / Service

  • A JSON file that describes:
    • Which container image(s) to use
    • How much CPU and memory to allocate
    • Which ports to expose
    • What environment variables to set
    • How containers should communicate with each other
    • Storage volumes to mount
    • Network settings
    • LIKE A BLUEPRINT
  • Task / service
    • you can create them by referencing the task definitions you created
    • Each cluster can run multiple tasks/services → Tasks/services are at the cluster level
    • Task: An instance of a task definition running on your cluster (the actual running containers
    • Service: A configuration that maintains a specified number of tasks running at all times (for long-running applications like web servers)
  • So when you’re in the AWS Management Console:
    • You create task definitions first (the blueprint)
    • Then you run tasks or create services based on those task definitions
  • Analogy
    • Task definition = Recipe
    • Task = Meal prepared from that recipe
    • Service = Chef who keeps making that meal whenever it’s consumed

Typical workflow

  1. You build your container image using Docker on your development machine or in a CI/CD pipeline
  2. You push that image to a container registry like ECR (Elastic Container Registry), Docker Hub, or another registry
  3. You define how to run your container using:
    • ECS: Task definitions and services
    • EKS: Kubernetes manifests (deployments, services, etc.)
  4. ECS/EKS pulls your image from the registry and runs it on the underlying infrastructure

Cluster

ECS/EKS manage clusters of infrastructure (the compute resources)

  • A collection of infrastructure resources (servers/compute), managed as a single unit
  • Can run many containers across multiple machines
  • Provides the environment where containers run
  • Handles scheduling, networking, and management
  • The relationship w/ containers is hierarchical:
    • A cluster contains many nodes (servers/instances)
    • Each node can run multiple containers
    • The cluster orchestrates how containers are distributed across nodes

Graph View

Backlinks