Auto Scaling allows up to scale up and down the EC2 instance depending on conditions that is being defined by the solutions architect
Use Case
- Medium Corp is an E-commerce organization in India and is hosted completely on AWS. In the past 6 months, the AWS bills have skyrocketed and CEO has asked you to reconsider your design so that is not only supports operations but also does it in cost effective way
- Since it is E-Commerce company targeting Indian consumers, the traffic patterns drastically during daytime and nighttime
How to Achieve ?
Can be achieved depending on the average load of the instance
Simple Scaling Policy
Base instance level: 2 years
Scalable
If average CPU utilization > 70% ; add 2 more instance
If average CPU utilization < 25% ; remove 2 more instance
Understanding Scalability
- Auto Scaling helps in achieving horizontal scalability of EC2 instances depending on the condition defined by the user
- To achieve the use case of Medium Corp, we can scale based on CPU utilization
- Nigh Time -> Less Traffic -> Low CPU Utilization -> Servers Scaled Down
- Day Time -> High Traffic -> High CPU Utilization -> Servers Scaled Up
Components of Auto Scaling
Auto Scaling has 2 important components:
- Launch Configuration
- Auto Scaling Groups
Launch Configuration (What)
- Name
- AMI
- Instance Type
- Security Groups
- IAM Role
- Etc…
Auto Scaling Groups (Where)
- Minimum and Maximum capacity
- VPC and Subnets
- Load Balancer
- Health Checks