etcdserver: rename to "heartbeat_send_failures_total"

Signed-off-by: Gyuho Lee <gyuhox@gmail.com>

etcdserver/metrics.go

@@ -42,10 +42,10 @@ var (
 		Name:      "leader_changes_seen_total",
 		Help:      "The number of leader changes seen.",
 	})
-	heartbeatFailures = prometheus.NewCounter(prometheus.CounterOpts{
+	heartbeatSendFailures = prometheus.NewCounter(prometheus.CounterOpts{
 		Namespace: "etcd",
 		Subsystem: "server",
-		Name:      "heartbeat_failures_total",
+		Name:      "heartbeat_send_failures_total",
 		Help:      "The total number of heartbeat send failures (likely overloaded from slow disk).",
 	})
 	slowApplies = prometheus.NewCounter(prometheus.CounterOpts{
@@ -97,7 +97,7 @@ func init() {
 	prometheus.MustRegister(hasLeader)
 	prometheus.MustRegister(isLeader)
 	prometheus.MustRegister(leaderChanges)
-	prometheus.MustRegister(heartbeatFailures)
+	prometheus.MustRegister(heartbeatSendFailures)
 	prometheus.MustRegister(slowApplies)
 	prometheus.MustRegister(proposalsCommitted)
 	prometheus.MustRegister(proposalsApplied)

etcdserver/raft.go

@@ -368,7 +368,7 @@ func (r *raftNode) processMessages(ms []raftpb.Message) []raftpb.Message {
 				// TODO: limit request rate.
 				if r.lg != nil {
 					r.lg.Warn(
-						"failed to send out heartbeat; took too long, server is overloaded likely from slow disk",
+						"leader failed to send out heartbeat on time; took too long, leader is overloaded likely from slow disk",
 						zap.Duration("heartbeat-interval", r.heartbeat),
 						zap.Duration("expected-duration", 2*r.heartbeat),
 						zap.Duration("exceeded-duration", exceed),
@@ -377,7 +377,7 @@ func (r *raftNode) processMessages(ms []raftpb.Message) []raftpb.Message {
 					plog.Warningf("failed to send out heartbeat on time (exceeded the %v timeout for %v)", r.heartbeat, exceed)
 					plog.Warningf("server is likely overloaded")
 				}
-				heartbeatFailures.Inc()
+				heartbeatSendFailures.Inc()
 			}
 		}
 	}