cluster upgrade fails with ClusterAPIReconcileTimeout when there is no change on both coe clustertemplate and capi clusterclass

[okozachenko1203]

In this case, coe cluster upgrade has no meaning on capi level, and capi controller will not increase the capi cluster's generation.
We have to filter this case and skip waiting

magnum-cluster-api/magnum_cluster_api/driver.py

Lines 307 to 319 in be7d18c

	def wait_capi_cluster_reconciliation_start(
	self, context, cluster: magnum_objects.Cluster, old_generation: int
	):
	"""Wait until the cluster's new generation is observed by capi-controller
	This means the cluster reconciliation has been started and the conditions has been updated.
	"""
	current_generation = resources.Cluster(
	context, self.k8s_api, cluster
	).get_observed_generation()
	if old_generation != current_generation:
	return
	raise exceptions.ClusterAPIReconcileTimeout()

[okozachenko1203]

The cluster upgrade will be used more frequently after this approach is implemented #324
And we can face this issue very soon with customers.