otlpgrpc: turn on round_robin LB policy and kuberesolver for resolving k8s service endpoints

CHANGELOG.md

@@ -6,6 +6,8 @@
 * [CHANGE] Index Cache: Multi level cache backfilling operation becomes async. Added `-blocks-storage.bucket-store.index-cache.multilevel.max-async-concurrency` and `-blocks-storage.bucket-store.index-cache.multilevel.max-async-buffer-size` configs and metric `cortex_store_multilevel_index_cache_backfill_dropped_items_total` for number of dropped items. #5661
 * [FEATURE] Ingester: Add per-tenant new metric `cortex_ingester_tsdb_data_replay_duration_seconds`. #5477
 * [FEATURE] Query Frontend/Scheduler: Add query priority support. #5605
+* [FEATURE] Tracing: Add `kuberesolver` to resolve endpoints address with `kubernetes://` prefix as Kubernetes service. #5731
+* [FEATURE] Tracing: Add `tracing.otel.round-robin` flag to use `round_robin` gRPC client side LB policy for sending OTLP traces. #5731
 * [ENHANCEMENT] Store Gateway: Added `-store-gateway.enabled-tenants` and `-store-gateway.disabled-tenants` to explicitly enable or disable store-gateway for specific tenants. #5638
 * [ENHANCEMENT] Compactor: Add new compactor metric `cortex_compactor_start_duration_seconds`. #5683
 * [ENHANCEMENT] Upgraded Docker base images to `alpine:3.18`. #5684

docs/configuration/config-file-reference.md

@@ -5024,6 +5024,12 @@ otel:
   # CLI flag: -tracing.otel.sample-ratio
   [sample_ratio: <float> | default = 0.001]
 
+  # If enabled, use round_robin gRPC load balancing policy. By default, use
+  # pick_first policy. For more details, please refer to
+  # https://github.com/grpc/grpc/blob/master/doc/load-balancing.md#load-balancing-policies.
+  # CLI flag: -tracing.otel.round-robin
+  [round_robin: <boolean> | default = false]
+
   # Enable TLS in the GRPC client. This flag needs to be enabled when any other
   # TLS flag is set. If set to false, insecure connection to gRPC server will be
   # used.

go.mod

@@ -80,6 +80,7 @@ require (
 	github.com/VictoriaMetrics/fastcache v1.12.1
 	github.com/cespare/xxhash/v2 v2.2.0
 	github.com/google/go-cmp v0.6.0
+	github.com/sercand/kuberesolver v2.4.0+incompatible
 	golang.org/x/exp v0.0.0-20231206192017-f3f8817b8deb
 	google.golang.org/protobuf v1.31.0
 )
@@ -196,7 +197,6 @@ require (
 	github.com/rs/cors v1.9.0 // indirect
 	github.com/rs/xid v1.5.0 // indirect
 	github.com/sean-/seed v0.0.0-20170313163322-e2103e2c3529 // indirect
-	github.com/sercand/kuberesolver v2.4.0+incompatible // indirect
 	github.com/shurcooL/httpfs v0.0.0-20230704072500-f1e31cf0ba5c // indirect
 	github.com/shurcooL/vfsgen v0.0.0-20200824052919-0d455de96546 // indirect
 	github.com/sirupsen/logrus v1.9.3 // indirect

pkg/tracing/tracing.go

@@ -8,6 +8,7 @@ import (
 
 	"github.com/go-kit/log/level"
 	"github.com/pkg/errors"
+	"github.com/sercand/kuberesolver"
 	"github.com/weaveworks/common/tracing"
 	"go.opentelemetry.io/contrib/propagators/aws/xray"
 	"google.golang.org/grpc/credentials"
@@ -41,6 +42,7 @@ type Otel struct {
 	OtlpEndpoint   string              `yaml:"otlp_endpoint" json:"otlp_endpoint"`
 	ExporterType   string              `yaml:"exporter_type" json:"exporter_type"`
 	SampleRatio    float64             `yaml:"sample_ratio" json:"sample_ratio"`
+	RoundRobin     bool                `yaml:"round_robin" json:"round_robin"`
 	TLSEnabled     bool                `yaml:"tls_enabled"`
 	TLS            tls.ClientConfig    `yaml:"tls"`
 	ExtraDetectors []resource.Detector `yaml:"-"`
@@ -55,6 +57,7 @@ func (c *Config) RegisterFlags(f *flag.FlagSet) {
 	f.StringVar(&c.Otel.OtlpEndpoint, p+".otel.otlp-endpoint", "", "otl collector endpoint that the driver will use to send spans.")
 	f.StringVar(&c.Otel.ExporterType, p+".otel.exporter-type", "", "enhance/modify traces/propagators for specific exporter. If empty, OTEL defaults will apply. Supported values are: `awsxray.`")
 	f.BoolVar(&c.Otel.TLSEnabled, p+".otel.tls-enabled", c.Otel.TLSEnabled, "Enable TLS in the GRPC client. This flag needs to be enabled when any other TLS flag is set. If set to false, insecure connection to gRPC server will be used.")
+	f.BoolVar(&c.Otel.RoundRobin, p+".otel.round-robin", false, "If enabled, use round_robin gRPC load balancing policy. By default, use pick_first policy. For more details, please refer to https://github.com/grpc/grpc/blob/master/doc/load-balancing.md#load-balancing-policies.")
 	c.Otel.TLS.RegisterFlagsWithPrefix(p+".otel.tls", f)
 }
 
@@ -91,15 +94,23 @@ func SetupTracing(ctx context.Context, name string, c Config) (func(context.Cont
 			level.Warn(util_log.Logger).Log("msg", "DEPRECATED: otel.otlp and otel.oltp both set, using otel.otlp because otel.oltp is deprecated")
 		}
 
+		endpoint := c.Otel.OtlpEndpoint
+		if (c.Otel.OtlpEndpoint == "") && (len(c.Otel.OltpEndpoint) > 0) {
+			level.Warn(util_log.Logger).Log("msg", "DEPRECATED: otel.oltp is deprecated use otel.otlp")
+			endpoint = c.Otel.OltpEndpoint
+		}
 		options := []otlptracegrpc.Option{
-			otlptracegrpc.WithEndpoint(c.Otel.OtlpEndpoint),
+			otlptracegrpc.WithEndpoint(endpoint),
+		}
+		// Following https://github.com/sercand/kuberesolver/blob/master/builder.go#L96.
+		if strings.HasPrefix(endpoint, "kubernetes://") {
+			// Registers the kuberesolver which resolves endpoint with prefix kubernetes://
+			// as kubernetes service endpoint addresses.
+			kuberesolver.RegisterInCluster()
 		}
 
-		if (c.Otel.OtlpEndpoint == "") && (len(c.Otel.OltpEndpoint) > 0) {
-			level.Warn(util_log.Logger).Log("msg", "DEPRECATED: otel.oltp is deprecated use otel.otlp")
-			options = []otlptracegrpc.Option{
-				otlptracegrpc.WithEndpoint(c.Otel.OltpEndpoint),
-			}
+		if c.Otel.RoundRobin {
+			options = append(options, otlptracegrpc.WithServiceConfig(`{"loadBalancingPolicy": "round_robin"}`))
 		}
 
 		if c.Otel.TLSEnabled {