Controlled Scaling in AWS
- This can be done by simple scaling or step Scaling
- Simple Scaling is a basic scaling policy where cloudwatch alarms trigger a single scaling action (add/remove instances)
- How it works:
- cloudwatch alarm is triggered (CPU > 80%)
- Autoscaling group performs a fixed adjustment (add 2 instances)
- A cool down period is applided to let the system stabilize before another action is taken
- Step scaling is more advanced scaling policy where action dependends on magnitude of alarm breach
-
How it works:
- cloudwatch alarm is triggered (CPU > 80%)
- Instead of one fixed action, multiple steps are defined
- if cpu is 80-90 add 1 instance
- if cpu > 90 add 2 instance
- Scaling continues until metric returns to normal range
-
Lets do simple and step scaling (Watch classroom recording)
