Auto Scaling Group

Auto Scaling Group (ASG) Overview

Auto Scaling Groups (ASGs) in AWS are designed to automatically adjust the number of EC2 instances in your infrastructure to match the load your application is experiencing. Here are the key points covered in the training:

Purpose of ASG:
- To automatically scale out (add EC2 instances) when there is an increased load.
- To scale in (remove EC2 instances) when there is a decreased load.
Size Variation:
- The number of instances in an ASG can vary over time based on defined parameters.
Minimum and Maximum Instances:
- You can set a minimum and maximum number of EC2 instances to ensure there's always a certain number of instances running.
Integration with Load Balancers:
- ASGs can be paired with load balancers to distribute traffic across instances.
- Unhealthy instances are terminated and replaced automatically.
Cost:
- ASGs themselves are free; you only pay for the EC2 instances and other resources used.
Capacity Settings:
- Minimum Capacity: The minimum number of instances in the ASG.
- Desired Capacity: The ideal number of instances in the ASG.
- Maximum Capacity: The maximum number of instances allowed in the ASG.

Untitled

Load Balancer (ELB) and ASG:
- ELB distributes traffic to instances in the ASG.
- ELB health checks can inform the ASG to terminate unhealthy instances.
Launch Templates:
- Replaces the deprecated launch configurations.
- Contains information on how to launch EC2 instances (AMI, instance type, user data, EBS volumes, security groups, etc.).

Untitled

Scaling Policies and CloudWatch Alarms:
- ASGs can scale based on CloudWatch alarms.
- Alarms can be set for metrics like average CPU utilization.
- Scaling policies can be created for scaling out (adding instances) or scaling in (removing instances).

Untitled

Example ASG Configuration

Attribute	Example Value
Minimum Capacity	2
Desired Capacity	4
Maximum Capacity	7
Launch Template	Specified AMI, instance type, etc.
Scaling Policies	Based on CloudWatch alarms

Conclusion

ASGs are essential for maintaining optimal performance and cost-efficiency in a cloud environment.
They work seamlessly with load balancers and CloudWatch alarms to ensure that the application can handle varying loads by adjusting the number of instances automatically.
The use of launch templates simplifies the process of defining how new instances should be configured.