AWS Classroom Series – 13/May/2021

Scaling in AWS

• Vertical Scaling:
  • Increasing resources to the EC2 instance i.e. changing the instance type
  • To achieve this we need to stop the ec2 instance and then resize
• Horizontal Scaling:
  • Increasing the EC2 instance which run the same application so that the load can be distributed.

Lab Setup

• To understand Horizontal Scaling and Elasticity lets create two AMIs
  • LAMP Stack on ubuntu:
    • Create a t2 micro ubuntu 18.04 ec2 instance

    sudo apt update
    sudo apt install apache2 -y 
    sudo apt install php libapache2-mod-php php-mysql php-cli -y
    # Create a file at /var/www/html/info.php with below content
    <?php
    phpinfo();
    ?>
    sudo apt install stress -y
    

    • Create AMI for this instance
  • Spring pet clinic on ubuntu
    • Steps for manual installation

    1  sudo apt update
    2  sudo apt install openjdk-11-jdk -y
    3  sudo mkdir /usr/share/springpetclinic
    4  wget https://referenceapplicationskhaja.s3-us-west-2.amazonaws.com/spring-petclinic-2.4.2.jar
    5  sudo cp spring-petclinic-2.4.2.jar /usr/share/springpetclinic/
    6  sudo adduser springbootuser
    7  sudo chown springbootuser:springbootuser /usr/share/springpetclinic/spring-petclinic-2.4.2.jar
    8  sudo chmod 500 /usr/share/springpetclinic/spring-petclinic-2.4.2.jar
    9  sudo vi /etc/systemd/system/springpetclinic.service
    10  sudo systemctl enable springpetclinic.service
    11  sudo systemctl start springpetclinic.service
    12  sudo systemctl status springpetclinic.service
    13 sudo apt install stress -y
    

    • Create an AMI from this ec2 instance
• When we run the AMI in an ec2 instance, users will access the application by giving public IP address
• In Reality we might be using domain names. (skip domain names)
• The application is run on webserver
• In AWS we have elastic load balancer which can distribute load across multiple ec2 instances and we have auto scaling group which can create multiple ec2 instances
  • depending on the static count
  • Scale out and scale in (increase or decrease the number of ec2 instances) based on ec2 metrics like CPU, network I/O
• To Create an autoscaling group, we need to have an AMI with application preinstalled or we can install after launch
• Lets explore autoscaling groups
• Auto Scaling groups requires one of the below
  • Launch Template
  • Launch Configuration
• In Launch Template/Launch Configuration we specify the details about
  • AMI to be used
  • Network in which ec2 instance has to be created
  • key pairs
  • security groups
  • instance types
• Then in Auto scaling group we specify
  • Conditions when the number of ec2 instances should increase or decrease
• To Effectively work with autoscaling groups we will associate
  • Elastic Load balancer
  • Application Load Balancer
  • Network Load Balancer
• For our lab purposes, to generate artificial load on the ec2 instances we will be stress tool.