Patterns and Recommendations

Congratulations! You’ve reached the final chapter of this book. Here, you’ll explore some patterns and recommendations for efficient Kubernetes autoscaling using KEDA and Karpenter. As I want you to continue practicing in this chapter, there will be some hands-on labs so you can see for yourself how to implement patterns such as scaling down to zero nodes and working with fault-tolerant workloads in Karpenter. These labs will help you maximize the cost optimization aspect of Karpenter.

As part of the hands-on labs, we’ll revisit Kubernetes constructs to improve your workload’s resiliency with settings such as terminationGracePeriodSeconds, leveraging preStop hooks, and using readiness probes, showing how these patterns plug into your autoscaling workflows to minimize downtime. You’ll also find guidance on topics such as integrating Spot capacity safely and increase your confidence to even use it in a production environment...