Exploring AWS !!

Day 12:

Autoscaling:

Scaling is adding/removing capacity/resource as needed.
Scale out is adding, scale in is removing.
Types: vertical and horizontal

Autoscaling is automatically scaling in/out.
Ensures correct number of instances are available to handle load.
Multi-Az EC2 instances provide high availability solution.
Autoscaling can dynamically increase/decrease capacity as needed.

Components:

Groups: EC2 instances are in groups so that they can be considered as a logical unit (for scaling and management). When we create groups we mention following attributes: min, max and desired number of instances.

Configuration Templates: used as config templates for EC2 instances. Launch template or launch configuration is also used.

Scaling options: Autoscaling provides several ways to scale group.
Manual scaling, Dynamic scaling, Scaling based on demand or schedule.

Scaling Policy:
Increase 2 instances at a time.
Decrease 1 instance at a time.
Alarm:
If CPU Utilization > 80% or more than 10 mins, ring bell.

Min capacity: 2
Desired capacity: 4
Max capacity: 10

Cool Down Period: ensures that auto scaling does not launch or terminate any instance until a specified period is completed. Scaling activity is suspended until cool-down period is in effect.

Step Scaling:
Alarm: CPU > 60% >60%/2
Action Add 2 instances >75%/3

The diagram represents the Lifecycle of Autoscaling.