Ingester: Defer converting from LabelPairs to Metric until flush (gra…

…fana#997)

* Defer converting from LabelPairs to Metric until needed

Since we average hundreds of samples to a chunk, it saves a lot of
work to leave label name/value sets as the slice they come in as,
rather than converting to the map type that Prometheus uses.

They are converted on flush, to avoid changing the chunk store.

* Store labelPairs in series sorted, for faster lookup
* Extract BenchmarkMetric as re-usable fixture for tests
* Add benchmark for Ingester Push()
* Copy name/value bytes, to avoid keeping the gRPC buffer alive

chunk_test.go

@@ -198,26 +198,7 @@ func TestChunksToMatrix(t *testing.T) {
 }
 
 func benchmarkChunk(now model.Time) Chunk {
-	// This is a real example from Kubernetes' embedded cAdvisor metrics, lightly obfuscated
-	return dummyChunkFor(now, model.Metric{
-		model.MetricNameLabel:              "container_cpu_usage_seconds_total",
-		"beta_kubernetes_io_arch":          "amd64",
-		"beta_kubernetes_io_instance_type": "c3.somesize",
-		"beta_kubernetes_io_os":            "linux",
-		"container_name":                   "some-name",
-		"cpu":                              "cpu01",
-		"failure_domain_beta_kubernetes_io_region": "somewhere-1",
-		"failure_domain_beta_kubernetes_io_zone":   "somewhere-1b",
-		"id":       "/kubepods/burstable/pod6e91c467-e4c5-11e7-ace3-0a97ed59c75e/a3c8498918bd6866349fed5a6f8c643b77c91836427fb6327913276ebc6bde28",
-		"image":    "registry/organisation/name@sha256:dca3d877a80008b45d71d7edc4fd2e44c0c8c8e7102ba5cbabec63a374d1d506",
-		"instance": "ip-111-11-1-11.ec2.internal",
-		"job":      "kubernetes-cadvisor",
-		"kubernetes_io_hostname": "ip-111-11-1-11",
-		"monitor":                "prod",
-		"name":                   "k8s_some-name_some-other-name-5j8s8_kube-system_6e91c467-e4c5-11e7-ace3-0a97ed59c75e_0",
-		"namespace":              "kube-system",
-		"pod_name":               "some-other-name-5j8s8",
-	})
+	return dummyChunkFor(now, BenchmarkMetric)
 }
 
 func BenchmarkEncode(b *testing.B) {

fixtures.go

@@ -0,0 +1,24 @@
+package chunk
+
+import "github.com/prometheus/common/model"
+
+// BenchmarkMetric is a real example from Kubernetes' embedded cAdvisor metrics, lightly obfuscated
+var BenchmarkMetric = model.Metric{
+	model.MetricNameLabel:              "container_cpu_usage_seconds_total",
+	"beta_kubernetes_io_arch":          "amd64",
+	"beta_kubernetes_io_instance_type": "c3.somesize",
+	"beta_kubernetes_io_os":            "linux",
+	"container_name":                   "some-name",
+	"cpu":                              "cpu01",
+	"failure_domain_beta_kubernetes_io_region": "somewhere-1",
+	"failure_domain_beta_kubernetes_io_zone":   "somewhere-1b",
+	"id":       "/kubepods/burstable/pod6e91c467-e4c5-11e7-ace3-0a97ed59c75e/a3c8498918bd6866349fed5a6f8c643b77c91836427fb6327913276ebc6bde28",
+	"image":    "registry/organisation/name@sha256:dca3d877a80008b45d71d7edc4fd2e44c0c8c8e7102ba5cbabec63a374d1d506",
+	"instance": "ip-111-11-1-11.ec2.internal",
+	"job":      "kubernetes-cadvisor",
+	"kubernetes_io_hostname": "ip-111-11-1-11",
+	"monitor":                "prod",
+	"name":                   "k8s_some-name_some-other-name-5j8s8_kube-system_6e91c467-e4c5-11e7-ace3-0a97ed59c75e_0",
+	"namespace":              "kube-system",
+	"pod_name":               "some-other-name-5j8s8",
+}