Home

Awesome

Microservices Orchestration on ECS

Arch

Complete Microservices Deploy and Orchestration on Amazon ECS using Terraform

Create clusters and services and pipelines on AWS using Terraform.

This is a incremental evolution from ecs-pipeline for big projects with more microservices and service discovery on AWS

AWS Stack Implementation:

Architecture

Arch

Deploy Pipeline

This demo project search for a buildspec.yml on root path from repository. You can see an example here.

<!-- ![Steps](.github/images/pipeline-demo.png) -->

How to Deploy

Edit AWS Configurations

Edit main.tf

# Customize your AWS Region
variable "aws_region" {
  description = "AWS Region for the VPC"
  default     = "us-east-1"
}

provider "aws" {
  region = var.aws_region
}

data "aws_caller_identity" "current" {}

Creating a cluster

Edit clusters.tf file to customize a cluster preferences. Give infos like ALB basic configurations, AZ's and etc.

module "cluster_example" {

    source              = "./modules/ecs"
    vpc_id              = module.vpc.vpc_id
    cluster_name        = var.cluster_name

    listener = {
        port     = 8080
        protocol = "HTTP"
        certificate_arn = ""
        ssl_policy      = "" // Default "ELBSecurityPolicy-TLS-1-1-2017-01"
    }

    availability_zones  = [
        module.vpc.public_subnet_1a,
        module.vpc.public_subnet_1b
    ]

}

Output for ecs

output "cluster_id" {}

output "alb" {}

output "listener" {}

Create an Service

Edit services.tf and customize an service configurations, like Github sources, containers preferences, alb route path and auto scale preferences.

module "service_whois" {
    source          = "./modules/service"
    vpc_id          = module.vpc.vpc_id
    region          = var.aws_region

    is_public       = true

    # Service name
    service_name        = "service-whois"
    service_base_path   = "/whois*"
    service_priority    = 400
    container_port      = 8080

    service_healthcheck = {
        healthy_threshold   = 3
        unhealthy_threshold = 10
        timeout             = 10
        interval            = 60
        matcher             = "200"
        path                = "/healthcheck"
        port                = 8080
    }

    # Cluster to deploy your service - see in clusters.tf
    cluster_name        = var.cluster_name
    cluster_id          = module.cluster_example.cluster_id
    cluster_listener    = module.cluster_example.listener
    cluster_mesh        = module.cluster_example.cluster_mesh

    cluster_service_discovery = module.cluster_example.cluster_service_discovery

    # Auto Scale Limits
    desired_tasks   = 2
    min_tasks       = 2
    max_tasks       = 10

    # Tasks CPU / Memory limits
    desired_task_cpu        = 256
    desired_task_mem        = 512

    # CPU metrics for Auto Scale
    cpu_to_scale_up         = 30
    cpu_to_scale_down       = 20
    cpu_verification_period = 60
    cpu_evaluation_periods  = 1

    # Pipeline Configuration
    build_image         = "aws/codebuild/docker:17.09.0"

    git_repository_owner    = "msfidelis"
    git_repository_name     = "microservice-nadave-whois"
    git_repository_branch   = "master"

    # AZ's
    availability_zones  = [
        module.vpc.public_subnet_1a,
        module.vpc.public_subnet_1b
    ]
}

Enable Container Insights

Just specify a value true on enable_container_insights parameter. (Default: false)

module "cluster_example" {
    source              = "./modules/ecs"

    vpc_id              = module.vpc.vpc_id
    cluster_name        = var.cluster_name

    // ...

    enable_container_insights   = true

    // ...
}

Using Fargate Spot (WAITING FOR TERRAFORM PROVIDER)

Just specify a value FARGATE_SPOT on service_launch_type parameter. (Default: FARGATE)

To change this value is necessary recreate a service. This is causes downtime on production.

How to Deploy

1) Github Access Token

export GITHUB_TOKEN=YOUR_TOKEN

2) Terraform

terraform init

terraform plan

terraform apply

References

Roadmap