As we move workloads to the cloud, we have the benefit of nearly unlimited capacity. In Chapter 3, Availability Patterns - Understanding Your Needs, we looked into how we could create an auto-scaling group. The resource we created used Spot Instances, a low-cost ephemeral compute capacity, to increase our number of instances:
scaling_adjustment = 4
adjustment_type = "ChangeInCapacity"
In the Persistence Patterns section of the book, we will dive into how to use our AWS metrics to choose what to measure. For now, we need to understand that whatever we choose as our default will at some point be suboptimal. Odds are that it will be on initial creation. We must constantly be validating our assumptions about what we've built.
Load testing with real data is one of the best ways to substantiate we are moving in the right direction...
