Metric for CloudProvider.Refresh() duration

This function can take an variable amount of time due to various
conditions (ie. many nodegroups changes causing forced refreshes,
caches time to live expiries, ...).

Monitoring that duration is useful to diagnose those variations,
and to uncover external issues (ie. throttling from cloud provider)
affecting cluster-autoscaler.

cluster-autoscaler/core/static_autoscaler.go

@@ -260,7 +260,9 @@ func (a *StaticAutoscaler) RunOnce(currentTime time.Time) errors.AutoscalerError
 	}
 
 	// Call CloudProvider.Refresh before any other calls to cloud provider.
+	refreshStart := time.Now()
 	err = a.AutoscalingContext.CloudProvider.Refresh()
+	metrics.UpdateDurationFromStart(metrics.CloudProviderRefresh, refreshStart)
 	if err != nil {
 		klog.Errorf("Failed to refresh cloud provider config: %v", err)
 		return errors.ToAutoscalerError(errors.CloudProviderError, err)

cluster-autoscaler/metrics/metrics.go

@@ -90,6 +90,7 @@ const (
 	FindUnneeded               FunctionLabel = "findUnneeded"
 	UpdateState                FunctionLabel = "updateClusterState"
 	FilterOutSchedulable       FunctionLabel = "filterOutSchedulable"
+	CloudProviderRefresh       FunctionLabel = "cloudProviderRefresh"
 	Main                       FunctionLabel = "main"
 	Poll                       FunctionLabel = "poll"
 	Reconfigure                FunctionLabel = "reconfigure"