I’m currently learning DevOps Beginners to Advanced with Projects on Udemy. Here are some notes on using AWS.

Original Application Stack

Project GitHub

• Nginx
• Apache
• Tomcat
• RabbitMQ
• Memcache
• Mysql

  

Migration Goals

• EC2
  • VMs for tomcat, rabbitmq, memcache, mysql
• ELB (Elastic Load Balancer)
  • Replaces nginx for load balancing
• Autoscaling
  • Automation for VM scaling
• S3/EFS
  • Shared storage
• Route 53
  • Private DNS service

Target Architecture

Flow of Execution

1. Create key pairs
2. Create security groups
   • Split into three groups:
     • LB (replaces nginx)
     • APP (for tomcat)
     • Backend (including rabbitmq, memcache, mysql)
3. Launch instances with user data

• Currently a semi-automated process
• Manually create instances and paste shell scripts for environment setup into userdata

4. Update IP to name mapping in Route 53
   • Set up an internal DNS for communication between instances using hostnames
5. Build the application from source code
   • This part is still semi-automated. Build the Java project on the local machine.
6. Upload to S3 bucket
   • Use AWS CLI to upload the built Java WAR file to the APP instance.
7. Download artifact to Tomcat EC2 instance
   • Previously, we used keys for S3 access. Here, instances connect to S3 using IAM roles.
     • Create a new S3 access role in IAM.
     • Attach the created role to the APP instance.
     • Use aws s3 ls to confirm successful access.
8. Set up ELB with HTTPS (certificate from Amazon Certificate Manager)
   • Create a target group, ensuring it points to port 8080 on the app.
   • Create an ELB with HTTP/HTTPS routing to the target group.
   • Purchase a domain and apply for an SSL certificate from AWS Certificate Manager.
   • In the secure listener, select the SSL certificate from ACM.
9. Map ELB endpoint to website name in DNS
   • At the DNS provider (in this case, GoDaddy), create a CNAME record that redirects to the AWS LB domain.
10. Verify
    • DNS settings may take some time to propagate.
    • You can directly access the LB’s domain to check if the APP is running on port 80.
11. Build an autoscaling group for the tomcat instance
    • Autoscaling involves three steps:
      • AMI (Amazon Machine Image)
        • Create an image from the current APP instance.
      • Launch template
        • Use the created AMI, and keep the security group the same as the original APP.
      • Autoscaling group
        • Attach it to the existing load balancer.
        • Choose the load balancer’s target group.
        • Set scaling policies based on CPU usage or network in/out.
        • Configure notifications.