It's possible to have the same pulumi stack to create a GKE cluster (

gcp.container.Cluster

) and a usable kubernetes provider (

kubernetes.Provider

) for that same cluster in the same run? I'm getting

error: kubernetes:<http://helm.sh/v4:Chart|helm.sh/v4:Chart> resource 'external-dns' has a problem: configured Kubernetes cluster is unreachable: unable to load Kubernetes client configuration from kubeconfig file. Make sure you have:

         • set up the provider as per <https://www.pulumi.com/registry/packages/kubernetes/installation-configuration/>

     invalid configuration: [context was not found for specified context: gke_xxxxxx_us-central1-a_main, cluster has no server defined]
    error: Program failed with an unhandled exception:

I can't find a way to make the kubernetes.Provider wait to until the cluster is actually created and the

~/.kube/config

updated. I have a

command.local.Command

that runs after the

gcp.container.Cluster

and updates the kubeconfig, But it seems that pulumi want to read the kubeconfig before it's updated. I want to do this in the same stack if possible as this is a stack that I have for testing so I want to created and destroy in the same day. (I already verified that if I comment the helm chart part, I can create the gke cluster and if I then uncomment the chart then the helm release is created since the kubeconfig is already updated in the second run).

Yes - it's 100% possible! https://github.com/pulumi/workshops/blob/main/gcloud-intro-kubernetes/index.ts

@stocky-restaurant-98004 I was alreeady using input/output for the

kubernetes.Provider

, the problem was that the

kubernetes.helm.v4.Chart

was not using that provider, I forgot to put

opts=pulumi.ResourceOptions(provider=kubernetes_provider)

so it was trying to use the default kubernetes provider.