I was part of cloud migration projects for one of our clients (a UK utility company), where we installed OpenShift clusters across different environments on top of the AWS cloud to deploy the code and run the DevOps pipelines.
We had five different OpenShift clusters viz; Development, Test, Pre-production, Production, and Automation. Each cluster would have different numbers of nodes, of which the master and infra nodes were constant across all environments (three nodes each) and a different number of worker nodes based on the workload on each cluster. So, the AWS cost was significantly high considering the number of nodes (AWS ec2-instances) we were working with across different clusters.