Add Neuron Scraper for scraping neuron monitor metrics

exporter/awsemfexporter/internal/appsignals/useragent.go

@@ -1,7 +1,7 @@
 // Copyright The OpenTelemetry Authors
 // SPDX-License-Identifier: Apache-2.0
 
-package appsignals
+package appsignals // import "github.com/open-telemetry/opentelemetry-collector-contrib/exporter/awsemfexporter/internal/appsignals"
 
 import (
 	"context"

exporter/awsemfexporter/metric_translator.go

@@ -132,7 +132,8 @@ func (mt metricTranslator) translateOTelToGroupedMetric(rm pmetric.ResourceMetri
 
 	if serviceName, ok := rm.Resource().Attributes().Get("service.name"); ok {
 		if strings.HasPrefix(serviceName.Str(), "containerInsightsKubeAPIServerScraper") ||
-			strings.HasPrefix(serviceName.Str(), "containerInsightsDCGMExporterScraper") {
+			strings.HasPrefix(serviceName.Str(), "containerInsightsDCGMExporterScraper") ||
+			strings.HasPrefix(serviceName.Str(), "containerInsightsNeuronMonitorScraper") {
 			// the prometheus metrics that come from the container insight receiver need to be clearly tagged as coming from container insights
 			metricReceiver = containerInsightsReceiver
 		}

exporter/awsemfexporter/metric_translator_test.go

@@ -282,6 +282,8 @@ func TestTranslateOtToGroupedMetric(t *testing.T) {
 	containerInsightMetric.Resource().Attributes().PutStr(conventions.AttributeServiceName, "containerInsightsKubeAPIServerScraper")
 	gpuMetric := createTestResourceMetricsHelper(defaultNumberOfTestMetrics + 1)
 	gpuMetric.Resource().Attributes().PutStr(conventions.AttributeServiceName, "containerInsightsDCGMExporterScraper")
+	neuronMetric := createTestResourceMetricsHelper(defaultNumberOfTestMetrics + 1)
+	neuronMetric.Resource().Attributes().PutStr(conventions.AttributeServiceName, "containerInsightsNeuronMonitorScraper")
 
 	counterSumMetrics := map[string]*metricInfo{
 		"spanCounter": {
@@ -390,6 +392,20 @@ func TestTranslateOtToGroupedMetric(t *testing.T) {
 			"myServiceNS/containerInsightsDCGMExporterScraper",
 			containerInsightsReceiver,
 		},
+		{
+			"neuron monitor receiver",
+			neuronMetric,
+			map[string]string{
+				"isItAnError": "false",
+				"spanName":    "testSpan",
+			},
+			map[string]string{
+				oTellibDimensionKey: "cloudwatch-lib",
+				"spanName":          "testSpan",
+			},
+			"myServiceNS/containerInsightsNeuronMonitorScraper",
+			containerInsightsReceiver,
+		},
 	}
 
 	for _, tc := range testCases {

internal/aws/containerinsight/const.go

@@ -157,11 +157,12 @@ const (
 	TypeContainerDiskIO    = "ContainerDiskIO"
 	// Special type for pause container
 	// because containerd does not set container name pause container name to POD like docker does.
-	TypeInfraContainer = "InfraContainer"
-	TypeGpuContainer   = "ContainerGPU"
-	TypeGpuPod         = "PodGPU"
-	TypeGpuNode        = "NodeGPU"
-	TypeGpuCluster     = "ClusterGPU"
+	TypeInfraContainer  = "InfraContainer"
+	TypeGpuContainer    = "ContainerGPU"
+	TypeGpuPod          = "PodGPU"
+	TypeGpuNode         = "NodeGPU"
+	TypeGpuCluster      = "ClusterGPU"
+	TypeNeuronContainer = "ContainerNeuron"
 
 	// unit
 	UnitBytes       = "Bytes"

receiver/awscontainerinsightreceiver/config.go

@@ -60,4 +60,8 @@ type Config struct {
 
 	// EnableAcceleratedComputeMetrics enabled features with accelerated compute resources where metrics are scraped from vendor specific sources
 	EnableAcceleratedComputeMetrics bool `mapstructure:"accelerated_compute_metrics"`
+
+	// EnableAwsNeuronMetrics toggles Neuron monitoring where metrics are scraped from neuron monitor
+	// The default value is false.
+	EnableAwsNeuronMetrics bool `mapstructure:"neuron_metrics"`
 }

receiver/awscontainerinsightreceiver/factory.go

@@ -42,6 +42,9 @@ const (
 
 	// Don't enable EKS control plane metrics by default
 	defaultEnableControlPlaneMetrics = false
+
+	// Don't enable Neuron metrics by default
+	defaultEnableAwsNeuronMetrics = false
 )
 
 // NewFactory creates a factory for AWS container insight receiver
@@ -64,6 +67,7 @@ func createDefaultConfig() component.Config {
 		ClusterName:                 defaultClusterName,
 		LeaderLockName:              defaultLeaderLockName,
 		EnableControlPlaneMetrics:   defaultEnableControlPlaneMetrics,
+		EnableAwsNeuronMetrics:      defaultEnableAwsNeuronMetrics,
 	}
 }
 

receiver/awscontainerinsightreceiver/internal/cadvisor/extractors/extractorhelpers.go

@@ -1,7 +1,7 @@
 // Copyright The OpenTelemetry Authors
 // SPDX-License-Identifier: Apache-2.0
 
-package extractors
+package extractors // import "github.com/open-telemetry/opentelemetry-collector-contrib/receiver/awscontainerinsightreceiver/internal/cadvisor/extractors"
 
 import (
 	"fmt"

receiver/awscontainerinsightreceiver/internal/gpu/dcgmscraper.go → receiver/awscontainerinsightreceiver/internal/gpu/dcgmscraper_config.go

@@ -1,12 +1,9 @@
 // Copyright The OpenTelemetry Authors
 // SPDX-License-Identifier: Apache-2.0
 
-package gpu
+package gpu // import "github.com/open-telemetry/opentelemetry-collector-contrib/receiver/awscontainerinsightreceiver/internal/gpu"
 
 import (
-	"context"
-	"errors"
-	"fmt"
 	"time"
 
 	configutil "github.com/prometheus/common/config"
@@ -15,14 +12,8 @@ import (
 	"github.com/prometheus/prometheus/discovery"
 	"github.com/prometheus/prometheus/discovery/kubernetes"
 	"github.com/prometheus/prometheus/model/relabel"
-	"go.opentelemetry.io/collector/component"
-	"go.opentelemetry.io/collector/consumer"
-	"go.opentelemetry.io/collector/pdata/pmetric"
-	"go.opentelemetry.io/collector/receiver"
-	"go.uber.org/zap"
 
 	ci "github.com/open-telemetry/opentelemetry-collector-contrib/internal/aws/containerinsight"
-	"github.com/open-telemetry/opentelemetry-collector-contrib/receiver/prometheusreceiver"
 )
 
 const (
@@ -33,76 +24,12 @@ const (
 	scraperK8sServiceSelector = "k8s-app=dcgm-exporter-service"
 )
 
-type DcgmScraper struct {
-	ctx                context.Context
-	settings           component.TelemetrySettings
-	host               component.Host
-	hostInfoProvider   hostInfoProvider
-	prometheusReceiver receiver.Metrics
-	k8sDecorator       Decorator
-	running            bool
-}
-
-type DcgmScraperOpts struct {
-	Ctx               context.Context
-	TelemetrySettings component.TelemetrySettings
-	Consumer          consumer.Metrics
-	Host              component.Host
-	HostInfoProvider  hostInfoProvider
-	K8sDecorator      Decorator
-	Logger            *zap.Logger
-}
-
 type hostInfoProvider interface {
 	GetClusterName() string
 	GetInstanceID() string
 }
 
-func NewDcgmScraper(opts DcgmScraperOpts) (*DcgmScraper, error) {
-	if opts.Consumer == nil {
-		return nil, errors.New("consumer cannot be nil")
-	}
-	if opts.Host == nil {
-		return nil, errors.New("host cannot be nil")
-	}
-	if opts.HostInfoProvider == nil {
-		return nil, errors.New("cluster name provider cannot be nil")
-	}
-
-	promConfig := prometheusreceiver.Config{
-		PrometheusConfig: &config.Config{
-			ScrapeConfigs: []*config.ScrapeConfig{getScraperConfig(opts.HostInfoProvider)},
-		},
-	}
-
-	params := receiver.CreateSettings{
-		TelemetrySettings: opts.TelemetrySettings,
-	}
-
-	decoConsumer := decorateConsumer{
-		containerOrchestrator: ci.EKS,
-		nextConsumer:          opts.Consumer,
-		k8sDecorator:          opts.K8sDecorator,
-		logger:                opts.Logger,
-	}
-
-	promFactory := prometheusreceiver.NewFactory()
-	promReceiver, err := promFactory.CreateMetricsReceiver(opts.Ctx, params, &promConfig, &decoConsumer)
-	if err != nil {
-		return nil, fmt.Errorf("failed to create prometheus receiver: %w", err)
-	}
-
-	return &DcgmScraper{
-		ctx:                opts.Ctx,
-		settings:           opts.TelemetrySettings,
-		host:               opts.Host,
-		hostInfoProvider:   opts.HostInfoProvider,
-		prometheusReceiver: promReceiver,
-		k8sDecorator:       opts.K8sDecorator,
-	}, nil
-}
-
-func getScraperConfig(hostInfoProvider hostInfoProvider) *config.ScrapeConfig {
+func GetScraperConfig(hostInfoProvider hostInfoProvider) *config.ScrapeConfig {
 	return &config.ScrapeConfig{
 		HTTPClientConfig: configutil.HTTPClientConfig{
 			TLSConfig: configutil.TLSConfig{
@@ -201,35 +128,3 @@ func getMetricRelabelConfig(hostInfoProvider hostInfoProvider) []*relabel.Config
 		},
 	}
 }
-
-func (ds *DcgmScraper) GetMetrics() []pmetric.Metrics {
-	// This method will never return metrics because the metrics are collected by the scraper.
-	// This method will ensure the scraper is running
-	if !ds.running {
-		ds.settings.Logger.Info("The scraper is not running, starting up the scraper")
-		err := ds.prometheusReceiver.Start(ds.ctx, ds.host)
-		if err != nil {
-			ds.settings.Logger.Error("Unable to start PrometheusReceiver", zap.Error(err))
-		}
-		ds.running = err == nil
-	}
-
-	return nil
-}
-
-func (ds *DcgmScraper) Shutdown() {
-	if ds.running {
-		err := ds.prometheusReceiver.Shutdown(ds.ctx)
-		if err != nil {
-			ds.settings.Logger.Error("Unable to shutdown PrometheusReceiver", zap.Error(err))
-		}
-		ds.running = false
-	}
-
-	if ds.k8sDecorator != nil {
-		err := ds.k8sDecorator.Shutdown()
-		if err != nil {
-			ds.settings.Logger.Error("Unable to shutdown K8sDecorator", zap.Error(err))
-		}
-	}
-}

receiver/awscontainerinsightreceiver/internal/gpu/dcgmscraper_test.go

@@ -21,7 +21,7 @@ import (
 
 	ci "github.com/open-telemetry/opentelemetry-collector-contrib/internal/aws/containerinsight"
 	"github.com/open-telemetry/opentelemetry-collector-contrib/receiver/awscontainerinsightreceiver/internal/mocks"
-	"github.com/open-telemetry/opentelemetry-collector-contrib/receiver/awscontainerinsightreceiver/internal/stores"
+	"github.com/open-telemetry/opentelemetry-collector-contrib/receiver/awscontainerinsightreceiver/internal/prometheusscraper"
 	"github.com/open-telemetry/opentelemetry-collector-contrib/receiver/prometheusreceiver"
 )
 
@@ -50,17 +50,6 @@ func (m mockHostInfoProvider) GetInstanceID() string {
 	return dummyInstanceID
 }
 
-type mockDecorator struct {
-}
-
-func (m mockDecorator) Decorate(metric stores.CIMetric) stores.CIMetric {
-	return metric
-}
-
-func (m mockDecorator) Shutdown() error {
-	return nil
-}
-
 type mockConsumer struct {
 	t        *testing.T
 	called   *bool
@@ -103,42 +92,6 @@ func (m mockConsumer) ConsumeMetrics(_ context.Context, md pmetric.Metrics) erro
 	return nil
 }
 
-func TestNewDcgmScraperBadInputs(t *testing.T) {
-	settings := componenttest.NewNopTelemetrySettings()
-	settings.Logger, _ = zap.NewDevelopment()
-
-	tests := []DcgmScraperOpts{
-		{
-			Ctx:               context.TODO(),
-			TelemetrySettings: settings,
-			Consumer:          nil,
-			Host:              componenttest.NewNopHost(),
-			HostInfoProvider:  mockHostInfoProvider{},
-		},
-		{
-			Ctx:               context.TODO(),
-			TelemetrySettings: settings,
-			Consumer:          mockConsumer{},
-			Host:              nil,
-			HostInfoProvider:  mockHostInfoProvider{},
-		},
-		{
-			Ctx:               context.TODO(),
-			TelemetrySettings: settings,
-			Consumer:          mockConsumer{},
-			Host:              componenttest.NewNopHost(),
-			HostInfoProvider:  nil,
-		},
-	}
-
-	for _, tt := range tests {
-		scraper, err := NewDcgmScraper(tt)
-
-		assert.Error(t, err)
-		assert.Nil(t, scraper)
-	}
-}
-
 func TestNewDcgmScraperEndToEnd(t *testing.T) {
 	expected := map[string]struct {
 		value  float64
@@ -182,16 +135,17 @@ func TestNewDcgmScraperEndToEnd(t *testing.T) {
 	settings := componenttest.NewNopTelemetrySettings()
 	settings.Logger, _ = zap.NewDevelopment()
 
-	scraper, err := NewDcgmScraper(DcgmScraperOpts{
+	scraper, err := prometheusscraper.NewSimplePrometheusScraper(prometheusscraper.SimplePrometheusScraperOpts{
 		Ctx:               context.TODO(),
 		TelemetrySettings: settings,
 		Consumer:          consumer,
 		Host:              componenttest.NewNopHost(),
 		HostInfoProvider:  mockHostInfoProvider{},
-		K8sDecorator:      mockDecorator{},
+		ScraperConfigs:    GetScraperConfig(mockHostInfoProvider{}),
+		Logger:            settings.Logger,
 	})
 	assert.NoError(t, err)
-	assert.Equal(t, mockHostInfoProvider{}, scraper.hostInfoProvider)
+	assert.Equal(t, mockHostInfoProvider{}, scraper.HostInfoProvider)
 
 	// build up a new PR
 	promFactory := prometheusreceiver.NewFactory()
@@ -207,7 +161,7 @@ func TestNewDcgmScraperEndToEnd(t *testing.T) {
 	mp, cfg, err := mocks.SetupMockPrometheus(targets...)
 	assert.NoError(t, err)
 
-	scrapeConfig := getScraperConfig(scraper.hostInfoProvider)
+	scrapeConfig := scraper.ScraperConfigs
 	scrapeConfig.ScrapeInterval = cfg.ScrapeConfigs[0].ScrapeInterval
 	scrapeConfig.ScrapeTimeout = cfg.ScrapeConfigs[0].ScrapeInterval
 	scrapeConfig.Scheme = "http"
@@ -238,9 +192,10 @@ func TestNewDcgmScraperEndToEnd(t *testing.T) {
 
 	// replace the prom receiver
 	params := receiver.CreateSettings{
-		TelemetrySettings: scraper.settings,
+		TelemetrySettings: scraper.Settings,
 	}
-	scraper.prometheusReceiver, err = promFactory.CreateMetricsReceiver(scraper.ctx, params, &promConfig, consumer)
+	scraper.PrometheusReceiver, err = promFactory.CreateMetricsReceiver(scraper.Ctx, params, &promConfig, consumer)
+
 	assert.NoError(t, err)
 	assert.NotNil(t, mp)
 	defer mp.Close()