Demo Cloud Solution

• Cloud example 1 - AWS / EC2 - just for self-learning
• Cloud example 2 - AWS / ECS Fargate - Main cloud solution

Prerequisites

• AWS account (Access Key ID and Secret Access Key required) with administrative permissions
• !!Only for EC2 example!! Generated SSH KeyPair and imported in AWS account, *.pem file saved on local disk (you will be prompted for file name and path)

EC2 deployment

• TerraformEC2 folder
• run terraform init / apply
• after successfull deployment, web application should be accessible via http://<public_ip>:8080

Front APP

• will be deployed automatically
• static HTML page
• exposed on port 8080
• doesn't communicate with backend DB for now

Backend APP just for learning - not in use

• POSTGRE SQL database
• exposed on port 5432

ECS deployment

• TerraformECSFargate folder
• My "Hi, Folks" web app doesn't use database, but there is an example how to create PostgreSQL database in AWS
• CI/CD workflow is done on GitHub side, described below

"Hi, Folks!" app - Service A / Service B

• two different revisions of the webapp used in EC2 example - service-a and service-b directories
• static HTML page served by nginx, containerized by Docker
• public Docker hub repository: seyhello/hi-folks-svc-a[b]
• CI workflow (defined in BuildServiceA[B]Container.yaml): docker image is built and pushed to the docker hub automatically when there are changes in service-a or service-b directories. "git push" starts the workflow
• CD workflow (defined in DeployServiceA[B].yaml): starts CI workflow and deploy the service to the AWS ECS cluster as Fargate service. It is ok as demonstration example because fixed variables are used in .yaml file. There would be better to use dynamic variables, but I didn't find out the way how to do it. Publish release on the git starts the workflow

Creating cluster

• just run terraform init / apply in the root directory
• you will be prompted for the VPC name (or project name, in this case)
• VPC, subnets, IGW, IAM role and ECS cluster is created

Managing service(s)

• To make it simple for you, there are two BASH scripts in the root project directory for managing services
• several resources are created
  • task definition for service
  • application load balancer
  • security groups
  • AWS CloudWatch log group and log stream for the service

add-service.sh

• script gets terraform output from cluster creation and service parameters specified in TFVARS file, and finally run terraform init / apply in its own directory and deploys the service to the cluster
• you must specify .tfvars file as required parameter, for example:

  ./add-services.sh services/service-a.tfvars
  

• examples for service A and service B are in services directory
• TFVARS file defines variables required for service creation - service name, docker repository, application port, instances count, number of CPUs and memory. File should look like this

  app_name       = "service-a"
  app_image      = "seyhello/hi-folks:service-a"
  app_port       = 80
  app_count      = 2
  fargate_cpu    = 1024
  fargate_memory = 2048
  

• you will be also prompted for the project name. This will be used for unique naming AWS resources
• Service DNS name is printed as terraform output

remove-service.sh

• run this script for removing deployed service
• you will be prompted for the service name
• script will run terraform destroy in the service directory and finally deletes the directory

Managing database(s)

• there are two BASH scripts in the root project directory for managing services

add-database.sh

• script gets terraform output from cluster/VPC creation, prompts user for additional database parameters, and finally run terraform init / apply in its own directory and deploys the database to the VPC
• database parameters required as user input are:

  Database instance ID
  Inside database name
  Username used for database operations
  Password for the user
  

remove-database.sh

• script for removing deployed database
• you will be prompted for the database instance id
• script will run terraform destroy in the database directory and finally deletes the directory