Reduce memory allocations in handlePacket

This converts the byte buffer to a string by casting and parses the
string by looping over it instead of calling Split.

This usage of unsafe is taken from the strings.Builder package here:
cb5c82bc3d/src/strings/builder.go (L47)

Signed-off-by: Clayton O'Neill <claytono@github.com>

bridge_test.go

@@ -315,3 +315,54 @@ func TestHandlePacket(t *testing.T) {
 		}
 	}
 }
+
+func BenchmarkUDPListener(b *testing.B) {
+	scenarios := []struct {
+		name    string
+		metrics [][]byte
+	}{
+		{
+			name:    "simple counter",
+			metrics: [][]byte{[]byte("counter:2|c")},
+		}, {
+			name:    "simple gauge",
+			metrics: [][]byte{[]byte("gauge:10|g")},
+		}, {
+			name:    "simple timing",
+			metrics: [][]byte{[]byte("timing:200|ms")},
+		}, {
+			name:    "simple histogram",
+			metrics: [][]byte{[]byte("histogram:200|h")},
+		}, {
+			name:    "simple distribution",
+			metrics: [][]byte{[]byte("distribution:200|d")},
+		}, {
+			name:    "simple_tags",
+			metrics: [][]byte{[]byte("simple_tags:100|c|#tag1:bar,tag2:baz")},
+		}, {
+			name:    "datadog tag extension with complex tags",
+			metrics: [][]byte{[]byte("foo:100|c|#09digits:0,tag.with.dots:1")},
+		}, {
+			name:    "datadog many tags",
+			metrics: [][]byte{[]byte("cpu_throttle_time_ms:1.1|ms|#action:test,application:testapp,application_component:testcomp,application_role:test_role,category:category,controller:controller,deployed_to:production,kube_deployment:deploy,kube_namespace:kube-production,method:get,version:rails5_2,status:200,status_range:2xx)}")},
+		}, {
+			name:    "datadog tag extension with sampling",
+			metrics: [][]byte{[]byte("foo:100|c|@0.1|#tag1:bar,#tag2:baz")},
+		}, {
+			name:    "combined multiline metrics",
+			metrics: [][]byte{[]byte("foo:200|ms:300|ms:5|c|@0.1:6|g\nbar:1|c:5|ms")},
+		},
+	}
+	for _, s := range scenarios {
+		l := &StatsDUDPListener{}
+		b.Run(s.name, func(b *testing.B) {
+			events := make(chan Events, len(s.metrics)*b.N)
+			defer close(events)
+			for i := 1; i <= b.N; i++ {
+				for _, m := range s.metrics {
+					l.handlePacket(m, events)
+				}
+			}
+		})
+	}
+}

exporter.go

@@ -26,6 +26,7 @@ import (
 	"strings"
 	"time"
 	"unicode/utf8"
+	"unsafe"
 
 	"github.com/prometheus/client_golang/prometheus"
 	"github.com/prometheus/common/log"
@@ -714,6 +715,7 @@ func parseDogStatsDTagsToLabels(component string) map[string]string {
 }
 
 func lineToEvents(line string) Events {
+	linesReceived.Inc()
 	events := Events{}
 	if line == "" {
 		return events
@@ -836,12 +838,26 @@ func (l *StatsDUDPListener) Listen(e chan<- Events) {
 
 func (l *StatsDUDPListener) handlePacket(packet []byte, e chan<- Events) {
 	udpPackets.Inc()
-	lines := strings.Split(string(packet), "\n")
+
 	events := Events{}
-	for _, line := range lines {
-		linesReceived.Inc()
+	// Convert the []byte array into a string w/o copying or allocating new
+	// memory, then walk the string looking for newlines to avoid memory
+	// allocation from Split.
+	p := *(*string)(unsafe.Pointer(&packet))
+	offset := 0
+	for i, c := range p {
+		if c == '\n' {
+			line := p[offset:i]
+			events = append(events, lineToEvents(line)...)
+			offset = i + 1
+		}
+	}
+
+	if offset < len(p) {
+		line := p[offset:]
 		events = append(events, lineToEvents(line)...)
 	}
+
 	e <- events
 }
 
@@ -884,7 +900,6 @@ func (l *StatsDTCPListener) handleConn(c *net.TCPConn, e chan<- Events) {
 			log.Debugf("Read %s failed: line too long", c.RemoteAddr())
 			break
 		}
-		linesReceived.Inc()
 		e <- lineToEvents(string(line))
 	}
 }
@@ -914,7 +929,6 @@ func (l *StatsDUnixgramListener) handlePacket(packet []byte, e chan<- Events) {
 	lines := strings.Split(string(packet), "\n")
 	events := Events{}
 	for _, line := range lines {
-		linesReceived.Inc()
 		events = append(events, lineToEvents(line)...)
 	}
 	e <- events