cloud-code-intellij
cloud-code-intellij copied to clipboard
Kubernetes Explorer does not show right contexts
The view has stale data, and forcing a refresh does not change it. I can't get it to show new clusters. There's also an issue with Cloud Run managed where it fails to load any GKE cluster saying there's a connection problem (not the case)
Cloud Code : Version 20.12.2 IntelliJ: 2020.3 Mac Os: 10.15.7
kubectl config get-contexts
CURRENT NAME CLUSTER AUTHINFO NAMESPACE
demo-cluster gke_vinnyc-demo-project_us-central1-c_demo-cluster gke_vinnyc-demo-project_us-central1-c_demo-cluster micronaut
* gke_vinnyc-demo-project_us-central1-c_cluster-1 gke_vinnyc-demo-project_us-central1-c_cluster-1 gke_vinnyc-demo-project_us-central1-c_cluster-1
jvm-tests gke_vinnyc-demo-project_us-central1-c_jvm-tests gke_vinnyc-demo-project_us-central1-c_jvm-tests default
Hi @viniciusccarvalho thanks for reporting your issue! Our Kubernetes explorer doesn't show context, it shows clusters, and it's a different thing - clusters may have a context, may not have a context (context deleted, but cluster is not deleted), and may have multiple contexts.
As we see from your screenshot, all clusters you have with your contexts are present in the view. You also have some other clusters that do not have contexts (i.e. gke_blue...
, ..tracing_cluster
, etc). Did you remove context for them but never removed the cluster itself? You can easily see which clusters do not have a context for them (we show context count next to the cluster name).
There's also an issue with Cloud Run managed where it fails to load any GKE cluster saying there's a connection problem (not the case)
Does this view show any Cloud Run services, i.e. managed ones, and fail to load Cloud Run Anthos (based on GKE clusters), or you're not able to see any Cloud Run services at all?
Hi @viniciusccarvalho thanks for reporting your issue! Our Kubernetes explorer doesn't show context, it shows clusters, and it's a different thing - clusters may have a context, may not have a context (context deleted, but cluster is not deleted), and may have multiple contexts.
As we see from your screenshot, all clusters you have with your contexts are present in the view. You also have some other clusters that do not have contexts (i.e.
gke_blue...
,..tracing_cluster
, etc). Did you remove context for them but never removed the cluster itself? You can easily see which clusters do not have a context for them (we show context count next to the cluster name).
Yes those clusters are long gone, its not intuitive how to remove them. I was expecting that cloud code would simply pull the context from my kubeconfig and use it.
So how do you remove those clusters? I was expecting a right click -> delete to be an option
The reason we do not include delete options for clusters and contexts is that it could easily go wrong and adding them back is not always easy. Delete operations on contexts and clusters are not expected to be a frequent action (i.e. it's not day to day operation). However, feel free to file an enhancement if you think delete functionality is something you're using often (for clusters, contexts, and actual resources such as pods/deployments etc.)
To remove a cluster, simply call
kubectl config delete-cluster NAME
@viniciusccarvalho with regards to the k8s issue: I believe, as @ivanporty mentioned, that Cloud Code is doing the right thing here. We display the information from your kubeconfig
file; if this is not the case for you please let us know. Regarding adding a delete
option, that is a feature request we can consider separately (though all of the caveats and dangers of this need to be considered).
For the Cloud Run service browser issue: would you mind providing a bit more detail (even in a new issue)? A screenshot of what you see in the IDE would help. Also - do you have Cloud Run services (that you can view in the Cloud web console)?
Delete has been implemented.