Getting started with Computing

2 chefs work when customers are few

Suddenly 100 customers arrive

The manager calls 3 more chefs

When customers leave, extra chefs go home

More website traffic → More EC2 instances

Less traffic → Fewer EC2 instances

Automatically adjusts the number of EC2 instances

Adds or removes servers based on demand

Works using defined scaling policies and metrics

How It Works

Monitors CPU usage, traffic, or custom metrics

Scales out when demand increases

Scales in when demand decreases

Keeps applications running smoothly

Ensures high availability and reliability

Handles sudden traffic spikes automatically

Reduces manual server management

Optimizes cost by using only required resources

Improves performance and resource efficiency

Horizontal scaling is more commonly used in AWS

Defines AMI, instance type, key pair, and

security groups Acts as a blueprint for launching instances

Key Components Included

Amazon Machine Image (AMI)

Instance type and storage settings

Network, security groups, and IAM role

User data scripts (startup configuration)

Required by Auto Scaling Groups

Ensures consistency across all EC2 instances

Prevents configuration mistakes

Enables faster and automated instance launches

Improves security by standardizing IAM roles and security groups

A logical group of EC2 instances managed as a single unit

All Auto Scaling actions happen inside an ASG

Uses Launch Templates or Launch Configurations

How ASG Works

Automatically launches new EC2 instances when demand increases.

Terminates instances when demand decreases

Works with scaling policies and CloudWatch metrics

Ensures high availability by running instances across multiple AZs

Automatically replaces unhealthy or failed instances

Maintains desired capacity at all times

Distributes traffic evenly when used with Load Balancers

Minimum Capacity

Minimum number of instances always running

Ensures basic availability

Prevents scaling below a safe limit

Maximum Capacity

Maximum number of instances allowed

Controls cost and resource usage

Prevents over-scaling during traffic spikes

Target number of instances ASG maintains

ASG scales in or out to reach this value

Can be manually or automatically adjusted

Rules that control when Auto Scaling adds or removes instances

Define scale-out and scale-in behavior

Triggered automatically based on conditions

How Scaling Decisions Are Made

Based on CPU utilization, network traffic, or custom metrics

Can also be time-based (scheduled scaling)

Uses CloudWatch metrics for monitoring

Target Tracking

Automatically keeps a chosen metric (like CPU usage) at a target values most commonly used

Step Scaling

Scales in or out in steps based on how much a metric crosses defined thresholds

Scheduled Scaling

Scales resources at predefined times based on expected traffic patterns.