AWS Classroom Series – 05/Dec/2020

AWS Autoscaling

• Autoscaling ensures application is able to handle load by increasing the number of ec2 instances when needed and decreasing the number of ec2 instances in the case of decreased load
• Autoscaling is used to ensure the desired number of ec2 instances are always running.
• Autoscaling Terms
  • Minimum Size
  • Desired capacity
  • Maximum Size
• Desired Capacity can be set to static number in such cases aws will ensure you will always have fixed set of ec2 instance running your application
• To make Desired Capacity Dynamic, We can specify any cloud watch metric such as cpu, network io, memory load etc to determine the desired capacity
• In the case of web servers & app servers as users increase or as load increases the stress will be more on CPU, so we will be using CPU metrics to determine the desired capacity
• Auto Scaling Components
  • Auto Scaling Groups:
    • We specify minimum, maximum & desired capacity
  • Configuration Templates:
    • We can use Launch templates or Launch configurations to determine the ami and other ec2 instance details
  • Scaling Options:
    • Dynamic
    • Fixed number
    • Manual Scaling

Launch Configuration (LC)

• LC is an instance configuration template that ASG uses to launch ec2 instance.
• Lets observe the different configuration details specified in creating a launch configuration
  • AMI-id
  • instance type
  • Additional Configuration
    • Spot instance requests
    • IAM Role
    • Monitoring
    • EBS optimized instance
    • User-Data
  • Volume
  • Security Groups
  • Key Pair

Launch Templates

• This is similar to Launch configuration. Launch template allows us to have multiple versions of the template.
• Launch template also gives option for dynamic values while creating
• Using Launch Template you can specify the same configuration as launch template, some of the values can be made dynamic and versioing is enabled.

Auto Scaling Groups (ASG)

• ASG contains collection of EC2 instances for the purpose of automatic scaling & management.
• ASG Group has features such as
  • Health check replacements
  • Scaling policies
• Size of ASG depends on number of instances in Desired Capacity which can be adjusted on demand, manually or by using automatic scaling

Scaling Options

• AWS EC2 Autoscaling provides the following scaling options
  • Maintains current instance levels at all times:
    • We configure ASG to maintain the specified number of running instance all the time.
  • Scale manually
  • Scale on a schedule:
    • Also referred as scheduled scaling.
  • Scale based on demand:
    • If your application needs a dynamic scaling policy
    • Scaling policy type:
      • Simple Scaling:
        • Increase or decrease the current capacity of the group based on single scaling adjustment
      • Step Scaling:
        • Increase or decrease the current capacity of the group based on set of scaling adjustement
      • Target tracking Scaling:
        • Increase or decrease the current capacity of group based on target value for a specific metric
    • Now configure Target tracking scaling and login into ec2 instance and run the stress (create artifical load on ec2 instances)
    • Now observe the autoscaling group for 10 mins and you should see the ec2 instances automatically added as cpu utilization moves beyond the configured number
• Exercise: Create an autoscaling group with launch template for running ec2 instances with
  • AMI: game of life
  • minimum size: 1
  • maximum size: 3
  • desired capacity: 1
  • Set Target tracking scaling policy to set at Average CPU utilization of 50%

Next Steps:

• Implementing a Loadbalancer
  • Layer 4
  • Layer 7
• Add a Route 53 route from DNS to gameoflife.