AWS Solution Architect Associate (SAA-C02) Review Material – ASG

General

• Adds or remove EC2 instances according to the load.
• Has minimum, maximum and desired number of instances.
• Can automatically register instances to ELB target group
• Used launch template or launch configuration to determine the instance type, volume, network (VPC & Subnet), security group, key pair of the instance to be launched.
• Scale-in = remove instances, Scale out = add instances
• To update ASG you need to provide new launch template or launch configuration.
• Terminates instances marked as unhealthy by ELB
• Cool Down Period – No instances will be launched or terminated. Default to 300 seconds.
• Has Lifecycle Hooks – Additional action can be performed before an instances is in service or terminated.
• Health Checks:
  • EC2
  • ELB

Types of Scaling

• Manual
  • By manually changing the minimum, maximum or desired capacity
• Dynamic (Scale based on metric)
  • Targeted
    • Scale-in when certain metric is above threshold, scale-out when below threshold.
    • example, when Average CPU Utilization is above a limit then increase instance.
    • Does not need to create a CloudWatch Alarm but can use CloudWatch Metrics
    • Can use predefined or customized metrics.
  • Simple
    • Required the use of CloudWatch Alarm
    • Based on CloudWatch Alarm can add/remove/set require instances.
    • Can specify the high / low of the alarm
    • Required instances can be a percentage of the group
  • Step
    • Required the use of CloudWatch Alarm
    • Similar to Simple but can have step adjustment e.g.
      • CPU > 30% add 1 instance
      • CPU > 50% add 2 instances
• Scheduled (Scale based on schedule)
  • Used when demand is known for a particular period in time.
• Predictive
  • Use machine learning to predict capacity requirement.
  • Good for:
    • Cyclical traffic, such as high use of resources during regular business hours and low use of resources during evenings and weekends
    • Recurring on-and-off workload patterns, such as batch processing, testing, or periodic data analysis
    • Applications that take a long time to initialize, causing a noticeable latency impact on application performance during scale-out events

Default Termination Policy

1. Determines the AZ with the most instances
2. Determine which instances is launched from the oldest launch template/configuration
3. Determine closest to the billing hour
4. Terminate in random.