terraform-provider-kubernetes VS aws-ami-gpu-monitoring

Now, to be honest, I'm not entirely sure/confident how this works. When I've used this kind of setup, I had two separate workspaces: one for setting up EKS and one for setting up Kubernetes within EKS. I'd apply the EKS workspace, first, then use its outputs for the Kubernete's workspace. You can see this pattern is specifically outlined in this EKS/k8s example. The Kubernetes provider docs also explicitly warns against creating the cluster in the same module as the Kubernetes provider. So it appears this may work, but it isn't recommended.

It is a pretty important distinction. Terraform and Kubernetes are fundamentally different in how they work. If you ever try to manage kubernetes state from terraform, it the differences become very obvious: https://github.com/hashicorp/terraform-provider-kubernetes/issues/1367

I'm using the terraform kubernetes-provider and I'd like to translate something like this kubectl command into TF:

After the resources be provisioned, you might want to run terraform apply -refresh-only to refresh your local state as the creation of some resource change the state of others within AWS. Also, state differences on metadata.resource_version of k8s resources almost always show up after an apply. This seems to be related to this issue.

You can open an issue here: https://github.com/hashicorp/terraform-provider-kubernetes/issues

We want to do that, but this issue has been a huge blocker for us. You might not hit it unless you’re using AKS, though.

From that repo, the only thing changed is the base AMI, which in this case an AMI tailored for accelerated hardware on EKS was used. The list of compatible AMIs for EKS can be obtained in this link updated regularly by AWS. Also, the AMI from AWS comes with SSM agent in it, so no need to change anything regarding that.