Autoscaling, cold storage, and other tricks for cost optimization
In Chapter 4, in the Autoscaling clusters and workloads section, we already discussed a key benefit and core capability that comes with Kubernetes. Several different tools and mechanisms are built into K8s to scale, such as configuring ReplicaSets (the number of instances per workload we want) or using observability data to drive automated scaling decisions using Horizontal Pod Autoscaler (HPA), Vertical Pod Autoscaler (VPA), or Kubernetes Event Driven Autoscaling (KEDA). There is a great free tutorial that walks through all the different options to autoscale on Kubernetes provided by Is It Observable. Here is the YouTube tutorial link, which also contains links to the GitHub tutorial: https://www.youtube.com/watch?v=qMP6tbKioLI.
The primary use case of autoscaling is to ensure that workloads have enough compute, memory, and storage to achieve certain availability goals. For our Financial One ACME company, this could...
