Skip to main content

Introduction to ECS and Fargate

1 - What is Amazon ECS?​

Amazon ECS (Elastic Container Service) is a fully managed container orchestration service that lets you run, stop, and manage Docker containers on a cluster.

2 - ECS vs EKS vs Fargate​

2.1 Comparison​

CriterionECSEKSFargate
TypeNative AWS orchestratorManaged KubernetesServerless compute
ComplexityLowHighVery low
PortabilityAWS onlyMulti-cloudAWS only
Infra managementEC2 or FargateEC2 or FargateNone
Learning curveGentleSteepVery gentle
Cost$$$$$

2.2 When to choose what?​

Choose ECS if:

  • You are 100% on AWS
  • You do not need Kubernetes
  • You want a simple solution

Choose EKS if:

  • You need Kubernetes
  • Multi-cloud portability is required
  • You have a team with K8s expertise

Choose Fargate if:

  • You do not want to manage infrastructure
  • Workloads with variable scaling
  • Maximum simplicity

3 - ECS architecture​

3.1 Main components​

3.2 Hierarchy of concepts​

ConceptDescription
ClusterLogical grouping of resources
Task DefinitionBlueprint for the containers (like a Dockerfile)
TaskA running instance of a Task Definition
ServiceMaintains N tasks running
Container InstanceEC2 with the ECS agent (EC2 launch type)

4 - Launch Types​

4.1 EC2 Launch Type​

Advantages:

  • Full control over the infrastructure
  • GPU and specialized instances
  • Reserved pricing possible
  • Access to the instances

Disadvantages:

  • Instance management
  • Capacity planning
  • OS patching

4.2 Fargate Launch Type​

Advantages:

  • Serverless - no infrastructure to manage
  • Pay-per-use
  • Instant scaling
  • Enhanced security (isolation)

Disadvantages:

  • Higher cost per task
  • No access to the instances
  • Resource limits

5 - Use cases​

5.1 Microservices​

5.2 Batch Processing​

# Task Definition pour batch
{
  "family": "batch-processor",
  "containerDefinitions": [
    {
      "name": "processor",
      "image": "batch-app:latest",
      "command": ["python", "process.py"]
    }
  ]
}

5.3 CI/CD Runners​

  • Jenkins/GitLab runners on ECS
  • Queue-based scaling
  • Isolation per container

6 - Pricing​

6.1 EC2 Launch Type​

ComponentCost
ECSFree
EC2 InstancesStandard EC2 price
Data TransferStandard price

6.2 Fargate​

ResourcePrice (eu-west-1)
vCPU~$0.04048/hour
Memory~$0.004445/GB/hour

Example: 1 vCPU, 2GB RAM, 24h

  • vCPU: 0.04048×24=0.04048 × 24 = 0.97
  • Memory: 0.004445×2×24=0.004445 × 2 × 24 = 0.21
  • Total: ~$1.18/day

6.3 Fargate Spot​

Up to 70% discount for workloads that tolerate interruptions.

7 - First ECS cluster​

7.1 Via the Console​

  1. Open the ECS console
  2. "Create Cluster"
  3. Choose the template (Networking only for Fargate)
  4. Name the cluster
  5. Configure VPC/Subnets

7.2 Via the CLI​

# Créer un cluster
aws ecs create-cluster \
    --cluster-name mon-cluster \
    --capacity-providers FARGATE FARGATE_SPOT \
    --default-capacity-provider-strategy \
        capacityProvider=FARGATE,weight=1 \
        capacityProvider=FARGATE_SPOT,weight=1

# Vérifier
aws ecs describe-clusters --clusters mon-cluster

7.3 Via CloudFormation​

AWSTemplateFormatVersion: '2010-09-09'

Resources:
  ECSCluster:
    Type: AWS::ECS::Cluster
    Properties:
      ClusterName: mon-cluster
      CapacityProviders:
        - FARGATE
        - FARGATE_SPOT
      DefaultCapacityProviderStrategy:
        - CapacityProvider: FARGATE
          Weight: 1
        - CapacityProvider: FARGATE_SPOT
          Weight: 1
      ClusterSettings:
        - Name: containerInsights
          Value: enabled
      Tags:
        - Key: Environment
          Value: Production

Summary​

In this chapter, we discovered:

  • ECS as AWS's container orchestrator
  • The differences between ECS, EKS and Fargate
  • The architecture and components of ECS
  • The launch types (EC2 vs Fargate)
  • Typical use cases
  • Pricing

Next step​

In the next chapter, we will dive deeper into the Fundamental concepts of ECS.

→ Next chapter: Fundamental concepts

← Back to the table of contents