📖 What is Amazon EC2 Auto Scaling?
EC2 Auto Scaling automatically adjusts the number of Amazon EC2 instances in your application based on defined policies and health checks. This ensures high availability and optimal performance by scaling resources in response to changing demand, reducing operational overhead.
"The exam emphasizes scaling policies: target tracking, step scaling, and simple scaling. Understand how Launch Templates and Launch Configurations are used. Distinguish between predictive scaling and reactive scaling. Be aware of instance warm-up considerations and lifecycle hooks."
📚 Certification: AWS Certified Solutions Architect - Associate (SAA-C03)
🔑 What are the Key Concepts of Amazon EC2 Auto Scaling?
- ▸ Scaling policies (Target Tracking, Step, Simple) define *how* Auto Scaling adjusts capacity based on metrics like CPU utilization or queue length.
- ▸ Launch Templates are preferred over Launch Configurations, offering versioning and more instance configuration options for consistent deployments.
- ▸ Lifecycle Hooks allow you to perform custom actions (e.g., software installation) *before* or *after* instances are launched or terminated.
- ▸ Predictive Scaling uses machine learning to forecast demand and proactively scale resources, while reactive scaling responds to current metrics.
- ▸ Instance warm-up considers the time it takes for an instance to become fully operational, preventing premature scaling decisions based on initial metrics.
🎯 How does Amazon EC2 Auto Scaling appear on the SAA-C03 Exam?
You may be asked to identify the best Auto Scaling policy to maintain consistent application performance during predictable peak hours, such as daily reports.
A scenario might describe an application experiencing intermittent spikes in traffic – determine which scaling policy would be most cost-effective and responsive.
Expect questions about troubleshooting Auto Scaling groups that are not scaling as expected, focusing on health checks, scaling policies, and instance limits.
❓ Frequently Asked Questions
When should I use Lifecycle Hooks?
Use Lifecycle Hooks when you need to run scripts or perform tasks on instances *before* they start serving traffic or *after* they are terminated, like updating software or deregistering from a load balancer.
What's the difference between predictive and reactive scaling, and when is each best used?
Predictive scaling anticipates future load, ideal for scheduled events. Reactive scaling responds to current load, better for unpredictable spikes. Predictive scaling requires sufficient historical data.
Can Auto Scaling scale across multiple Availability Zones?
Yes, Auto Scaling is designed to distribute instances across multiple Availability Zones for high availability. You configure the desired number of subnets per AZ within the Auto Scaling group.