Stale APIService Causing API Discovery Failures while running kubectl

Problem

Clients such as kubectl (v1.22+), client-go, and Dynatrace components experience API discovery failures, with repeated error messages like:

couldn't get resource list for packages.operators.coreos.com/v1: the server is currently unable to handle the request

Environment

• Platform9 Managed Kubernetes
• kubernetes v1.23 and Higher

Cause

A stale/deprecated APIService object (v1.packages.operators.coreos.com) still exists in the cluster and references a Kubernetes Service (v1-packages-operators-coreos-com) in the pf9-olm namespace, which no longer exists. This results in a ServiceNotFound condition, which newer clients treat as a hard error during API discovery.

Resolution

Backup the APIService definition (for audit purposes):

kubectl get apiservice v1.packages.operators.coreos.com -o yaml > broken-apiservice-backup.yaml

Delete the stale/deprecated APIService:

kubectl delete apiservice v1.packages.operators.coreos.com

Additional Information

The issue is not observed in kubectl versions older than v1.20.1 because older versions of kubectl and client-go handle unavailable or misconfigured APIService resources more gracefully—typically by logging warnings instead of failing or blocking discovery entirely.