Exactly, now you have people making an EKS cluster and then deploying a vendor supplied Helm chart to it. When something breaks they literally have no idea how to start fixing it. It’s deceptively “easy”