One of the great promises of cloud computing is that it will let you:
AWS keeps expanding its AWS Auto Scaling service, which can manage these two requirements across an enormous range of resource types, including ECS services, Lambda functions and RDS clusters.
For each resource that you put under the control of AWS Auto Scaling, you can choose whether to maximize availability or savings. You can also choose to adopt a balanced strategy, somewhere between these two extremes.
To decide when to scale resources up or down, AWS Auto Scaling tracks metrics like instance utilization rates. It uses either currently observed metrics (Dynamic Scaling) or future metrics predicted by an artificial intelligence that has observed your previous metrics (Predictive Scaling).
You set a Minimum Capacity and Maximum Capacity, which act as guard rails on how high or low AWS Auto Scaling can rescale your resources.
So far, so good.
But what if you are launching a product next week, and demand will hopefully be five times greater than anything you’ve previously seen?
What if the actual surge in demand is 10x? Or just 2x?
What if the actual surge ends up lasting twice as long as predicted? Or half as long?
Autopilot is great, but in situations like this you want a quick way to be able to grab the wheel.
Use it to instantly reset the Minimum Capacity and Maximum Capacity on your scalable targets. Trigger it using a schedule, or in response to a CloudWatch Event, or by manually clicking the Run Now button.