fix(kubernetes): sleep before GET request in WaitForKubernetesClusterState (#275)

Sleep before `GET` request in `WaitForKubernetesClusterState` to avoid getting stale state in some scenarios.
This will also align function behavior with other services.

Author: peknur

CHANGELOG.md

@@ -7,6 +7,9 @@ See updating [Changelog example here](https://keepachangelog.com/en/1.0.0/)
 ## Added
 - kubernetes: `WaitForKubernetesNodeGroupState` method for waiting the node group to achieve a desired state
 
+### Changed
+- kubernetes: sleep before `GET` request in `WaitForKubernetesClusterState`
+
 ## [6.8.2]
 ### Added
 - account: `NetworkPeerings`, `NTPExcessGiB`, `StorageMaxIOPS`, and `LoadBalancers` fields to the `ResourceLimits` struct.

upcloud/service/kubernetes.go

@@ -82,6 +82,7 @@ func (s *Service) WaitForKubernetesClusterState(ctx context.Context, r *request.
 
 	for {
 		attempts++
+		time.Sleep(sleepDuration)
 
 		details, err := s.GetKubernetesCluster(ctx, &request.GetKubernetesClusterRequest{
 			UUID: r.UUID,
@@ -100,8 +101,6 @@ func (s *Service) WaitForKubernetesClusterState(ctx context.Context, r *request.
 			return details, nil
 		}
 
-		time.Sleep(sleepDuration)
-
 		if time.Duration(attempts)*sleepDuration >= r.Timeout {
 			return nil, fmt.Errorf("timeout reached while waiting for Kubernetes cluster to enter state \"%s\"", r.DesiredState)
 		}