Workload Autoscaling Operations

Up until this point, we’ve been exploring multiple ways to make your workloads more efficient in Kubernetes. In other words, we’ve discussed how to ensure your workloads are not wasting the resources available in the cluster. You’ve learned that to have consistent results, you need to be intentional about the number of CPU and memory requests your pods have. By doing so, you can then make use of tools such as HPA, VPA, and KEDA to adjust the number of replicas based on utilization or events that impact the performance of your workloads.

However, how do you know you’re doing proper right-sizing for your workloads? How do you keep doing it continuously, as it’s not a one-time task? And how do you know why your workloads might not be scaling? Well, you need to learn how to do proper troubleshooting, read logs, interpret metrics, and watch for events in the tools you’re using.

In this chapter, we’...