Скачать или смотреть AWS Auto Scaling: Dynamic Scaling Policy, Predictive Scaling, and Scheduled Scaling Strategies

Amazon EC2 Auto Scaling offers several dynamic scaling policy options:

Target Tracking Scaling: This policy adjusts the group's capacity based on an Amazon CloudWatch metric and a specified target value. Think of it like your home thermostat: you set a desired temperature, and the thermostat handles the rest.

Step Scaling: Step scaling changes the group's capacity by applying step adjustments that vary according to the size of the alarm breach. This approach allows for more granular control over scaling actions.

Simple Scaling: Simple scaling involves making capacity adjustments using a single scaling adjustment, with a cooldown period between each scaling operation.

If your scaling needs are tied to a metric that changes in proportion to the number of instances in your Auto Scaling group, it's advisable to use target tracking scaling policies. In contrast, for other scenarios, step scaling policies are recommended.

Target tracking policies provide dynamic scaling that aligns with the actual load on your application, ensuring immediate capacity adjustments in response to load fluctuations. Moreover, this policy can adapt to changes over time, such as seasonal variations.

By default, new Auto Scaling groups start without any scaling policies. To enable automatic scaling, you can either set up scheduled scaling or predictive scaling when you don't employ any form of dynamic scaling.

Typically, when your workload experiences predictable traffic surges and your applications have extended initialization times, it's advisable to explore the advantages of predictive scaling. Predictive scaling offers the advantage of proactively adjusting capacity in anticipation of expected load increases, contrasting with dynamic scaling, which reacts to changes as they occur. This proactive approach not only accelerates scaling but can also result in cost savings on your EC2 expenses by preventing the need for excessive capacity provisioning.

Scheduled scaling enables you to establish a custom scaling timetable that aligns with anticipated fluctuations in load. For instance, if you consistently observe a weekly traffic pattern where your web application experiences a surge in activity on Wednesdays, maintains high levels on Thursdays, and tapers off on Fridays, you can arrange Amazon EC2 Auto Scaling to boost capacity on Wednesdays and reduce it on Fridays according to your specific schedule.



Reference:
https://docs.aws.amazon.com/
https://aws.amazon.com/whitepapers/
  / mraviteja9949  


Instagram:   / raviteja_mureboina  


My book(Azure Security): https://a.co/d/iGznse6 #ravitejamureboina