cluster-api: Timeout for clusterctl

User Story

As a user of clusterctl client or cli utility, it would greatly benefit me if I could specify a timeout when getting kubeconfig for the cluster:

# clusterctl get kubeconfig foo --kubeconfig-context bar --timeout 10s

or

kubeconfig, err := c.GetKubeconfig(&client.GetKubeconfigOptions{
	ParentKubeconfigPath:    f,
	ParentKubeconfigContext: parentContext,
	ManagedClusterNamespace: ref.Namespace,
	ManagedClusterName:      ref.Name,
	Timeout: "10s" // or time.Second * 10
})

.

Detailed Description

The current timeout is 300 seconds. If the management cluster is not reachable, clusterctl will fail after 10 retries to reach the cluster with a 30-second timeout per each try ( controller-runtime client default behavior). When running this command as an interactive user, this seems ok, because you can interrupt it with CTRL+C. However, when done as part of the CI/CD pipeline or another non-interactive script, we have to wait for 300s to understand that cluster is not reachable.

It would be a great help if we could specify this timeout. I think it is easy to implement by using different context here: https://github.com/kubernetes-sigs/cluster-api/blob/master/cmd/clusterctl/client/cluster/workload_cluster.go#L53

func (p *workloadCluster) GetKubeconfig(workloadClusterName string, namespace string, timeout time.Duration) (string, error) {
...
	ctxWithTimeout, cancel := context.WithTimeout(ctx, timeout)
	defer cancel()
	dataBytes, err := utilkubeconfig.FromSecret(ctxWithTimeout, cs, obj)
...
}

/kind feature /area clusterctl