An interesting lab for an overengineered solution from AWS for Kubernetes workloads right sizing.

Should you implement it this way? I don't know. But maybe, you want to play with GitOps, AWS Bedrock and all that stuff.

Also, it's funny how they say in the beginning that having VPA and Goldilocks inside a cluster is an overhead and additional management burden and then propose to create a cluster in GHA runtime and use generative AI to address that.

#aws #kubernetes