metrics: update MCDRebootErr & MCDPivotErr to add correct node label

pkg/daemon/metrics.go

@@ -39,7 +39,7 @@ var (
 		prometheus.GaugeOpts{
 			Name: "mcd_pivot_err",
 			Help: "errors encountered during pivot",
-		}, []string{"pivot_target", "err"})
+		}, []string{"node", "pivot_target", "err"})
 
 	// MCDState is state of mcd for indicated node (ex: degraded)
 	MCDState = prometheus.NewGaugeVec(
@@ -59,7 +59,7 @@ var (
 	MCDRebootErr = prometheus.NewGaugeVec(
 		prometheus.GaugeOpts{
 			Name: "mcd_reboot_err",
-		}, []string{"message", "err"})
+		}, []string{"node", "message", "err"})
 
 	// MCDUpdateState logs completed update or error
 	MCDUpdateState = prometheus.NewGaugeVec(
@@ -88,9 +88,9 @@ func registerMCDMetrics() error {
 		}
 	}
 
-	MCDPivotErr.WithLabelValues("", "").Set(0)
+	MCDPivotErr.WithLabelValues("", "", "").Set(0)
 	KubeletHealthState.Set(0)
-	MCDRebootErr.WithLabelValues("", "").Set(0)
+	MCDRebootErr.WithLabelValues("", "", "").Set(0)
 	MCDUpdateState.WithLabelValues("", "").Set(0)
 
 	return nil

pkg/daemon/update.go

@@ -1614,13 +1614,13 @@ func (dn *Daemon) reboot(rationale string) error {
 	// either, we just have one for the MCD itself.
 	if err := rebootCmd.Run(); err != nil {
 		dn.logSystem("failed to run reboot: %v", err)
-		MCDRebootErr.WithLabelValues("failed to run reboot", err.Error()).SetToCurrentTime()
+		MCDRebootErr.WithLabelValues(dn.node.Name, "failed to run reboot", err.Error()).SetToCurrentTime()
 	}
 
 	// wait to be killed via SIGTERM from the kubelet shutting down
 	time.Sleep(defaultRebootTimeout)
 
 	// if everything went well, this should be unreachable.
-	MCDRebootErr.WithLabelValues("reboot failed", "this error should be unreachable, something is seriously wrong").SetToCurrentTime()
+	MCDRebootErr.WithLabelValues(dn.node.Name, "reboot failed", "this error should be unreachable, something is seriously wrong").SetToCurrentTime()
 	return fmt.Errorf("reboot failed; this error should be unreachable, something is seriously wrong")
 }