In HyperShift 4.14, the konnectivity server is run inside the kube-apiserver pod. When this pod is deleted for any reason, the konnectivity server container can drop before the rest of the pod terminates, which can cause network connections to drop. The following preStop definition can be added to the container to ensure it stays alive long enough for the rest of the pod to clean up.

lifecycle:
  preStop:
    exec:
      command:
        - /bin/sh
        - -c
        - sleep 70

4.14

1.
2.
3.