count 429s in elasticsearch output

CHANGELOG.asciidoc

@@ -85,6 +85,7 @@ https://github.com/elastic/beats/compare/v6.4.0...master[Check the HEAD diff]
 - Make kubernetes autodiscover ignore events with empty container IDs {pull}7971[7971]
 - Add DNS processor with support for performing reverse lookups on IP addresses. {issue}7770[7770]
 - Implement CheckConfig in RunnerFactory to make autodiscover check configs {pull}7961[7961]
+- Count HTTP 429 responses in the elasticsearch output {pull}8056[8056]
 
 *Auditbeat*
 

libbeat/outputs/elasticsearch/client.go

@@ -112,6 +112,7 @@ type bulkResultStats struct {
 	duplicates   int // number of events failed with `create` due to ID already being indexed
 	fails        int // number of failed events (can be retried)
 	nonIndexable int // number of failed events (not indexable -> must be dropped)
+	tooMany      int // number of events receiving HTTP 429 Too Many Requests
 }
 
 var (
@@ -343,6 +344,7 @@ func (client *Client) publishEvents(
 		st.Failed(failed)
 		st.Dropped(dropped)
 		st.Duplicate(duplicates)
+		st.ErrTooMany(stats.tooMany)
 	}
 
 	if failed > 0 {
@@ -515,11 +517,15 @@ func bulkCollectPublishFails(
 			continue // ok
 		}
 
-		if status < 500 && status != 429 {
-			// hard failure, don't collect
-			logp.Warn("Cannot index event %#v (status=%v): %s", data[i], status, msg)
-			stats.nonIndexable++
-			continue
+		if status < 500 {
+			if status == http.StatusTooManyRequests {
+				stats.tooMany++
+			} else {
+				// hard failure, don't collect
+				logp.Warn("Cannot index event %#v (status=%v): %s", data[i], status, msg)
+				stats.nonIndexable++
+				continue
+			}
 		}
 
 		debugf("Bulk item insert failed (i=%v, status=%v): %s", i, status, msg)

libbeat/outputs/metrics.go

@@ -33,6 +33,7 @@ type Stats struct {
 	active     *monitoring.Uint // events sent and waiting for ACK/fail from output
 	duplicates *monitoring.Uint // events sent and waiting for ACK/fail from output
 	dropped    *monitoring.Uint // total number of invalid events dropped by the output
+	tooMany    *monitoring.Uint // total number of too many requests replies from output
 
 	//
 	// Output network connection stats
@@ -56,6 +57,7 @@ func NewStats(reg *monitoring.Registry) *Stats {
 		dropped:    monitoring.NewUint(reg, "events.dropped"),
 		duplicates: monitoring.NewUint(reg, "events.duplicates"),
 		active:     monitoring.NewUint(reg, "events.active"),
+		tooMany:    monitoring.NewUint(reg, "events.toomany"),
 
 		writeBytes:  monitoring.NewUint(reg, "write.bytes"),
 		writeErrors: monitoring.NewUint(reg, "write.errors"),
@@ -117,6 +119,13 @@ func (s *Stats) Cancelled(n int) {
 	}
 }
 
+// ErrTooMany updates the number of Too Many Requests responses reported by the output.
+func (s *Stats) ErrTooMany(n int) {
+	if s != nil {
+		s.tooMany.Add(uint64(n))
+	}
+}
+
 // WriteError increases the write I/O error metrics.
 func (s *Stats) WriteError(err error) {
 	if s != nil {

libbeat/outputs/observer.go

@@ -30,6 +30,7 @@ type Observer interface {
 	WriteBytes(int)   // report number of bytes being written
 	ReadError(error)  // report an I/O error on read
 	ReadBytes(int)    // report number of bytes being read
+	ErrTooMany(int)   // report too many requests response
 }
 
 type emptyObserver struct{}
@@ -51,3 +52,4 @@ func (*emptyObserver) WriteError(error) {}
 func (*emptyObserver) WriteBytes(int)   {}
 func (*emptyObserver) ReadError(error)  {}
 func (*emptyObserver) ReadBytes(int)    {}
+func (*emptyObserver) ErrTooMany(int)   {}