metrics

Documentation

Index

• Constants
• Variables
• func ComputeReconcilerNameType(nn types.NamespacedName) (reconcilerName, reconcilerType string)
• func RecordCRDCount(ctx context.Context, nn types.NamespacedName, count int64)
• func RecordClusterScopedResourceCount(ctx context.Context, nn types.NamespacedName, count int64)
• func RecordKCCResourceCount(ctx context.Context, nn types.NamespacedName, count int64)
• func RecordNamespaceCount(ctx context.Context, nn types.NamespacedName, count int64)
• func RecordPipelineError(ctx context.Context, nn types.NamespacedName, component string, hasErr bool)
• func RecordReadyResourceCount(ctx context.Context, nn types.NamespacedName, count int64)
• func RecordReconcileDuration(ctx context.Context, stallStatus string, startTime time.Time)
• func RecordResourceCount(ctx context.Context, nn types.NamespacedName, count int64)
• func RecordResourceGroupTotal(ctx context.Context, count int64)
• func RegisterOCAgentExporter() (*ocagent.Exporter, error)
• func RegisterReconcilerMetricsViews() error

Constants

const (
	// NsReconcilerPrefix is the prefix used for all Namespace reconcilers.
	NsReconcilerPrefix = "ns-reconciler"
	// RootReconcilerPrefix is the prefix usef for all Root reconcilers.
	RootReconcilerPrefix = "root-reconciler"
	// RepoSyncName is the expected name of any RepoSync CR.
	RepoSyncName = "repo-sync"
	// RootSyncName is the expected name of any RootSync CR.
	RootSyncName = "root-sync"

	//nolint:gosec // ignore the false-positive alert for G101: Potential hardcoded credentials
	// CMSNamespace is the name of the Config Sync controller's namespace
	CMSNamespace = "config-management-system"
)

Variables

var (
	// ReconcileDuration tracks the time duration in seconds of reconciling
	// a ResourceGroup CR by the ResourceGroup controller.
	// label `reason`: the `Reason` field of the `Stalled` condition in a ResourceGroup CR.
	// reason can be: StartReconciling, FinishReconciling, ComponentFailed, ExceedTimeout.
	// This metric should be updated in the ResourceGroup controller.
	ReconcileDuration = stats.Float64(
		"rg_reconcile_duration_seconds",
		"Time duration in seconds of reconciling a ResourceGroup CR by the ResourceGroup controller",
		stats.UnitSeconds)

	// ResourceGroupTotal tracks the total number of ResourceGroup CRs in a cluster.
	// This metric should be updated in the Root controller.
	ResourceGroupTotal = stats.Int64(
		"resource_group_total",
		"Total number of ResourceGroup CRs in a cluster",
		stats.UnitDimensionless)

	// ResourceCount tracks the number of resources in a ResourceGroup CR.
	// This metric should be updated in the Root controller.
	ResourceCount = stats.Int64(
		"resource_count",
		"The number of resources in a ResourceGroup CR",
		stats.UnitDimensionless)

	// ReadyResourceCount tracks the number of resources with Current status in a ResourceGroup CR.
	// This metric should be updated in the ResourceGroup controller.
	ReadyResourceCount = stats.Int64(
		"ready_resource_count",
		"The number of resources with Current status in a ResourceGroup CR",
		stats.UnitDimensionless)

	// KCCResourceCount tracks the number of KCC resources in a ResourceGroup CR.
	// This metric should be updated in the ResourceGroup controller.
	KCCResourceCount = stats.Int64(
		"kcc_resource_count",
		"The number of KCC resources in a ResourceGroup CR",
		stats.UnitDimensionless)

	// NamespaceCount tracks the number of resource namespaces in a ResourceGroup CR.
	// This metric should be updated in the Root controller.
	NamespaceCount = stats.Int64(
		"resource_ns_count",
		"The number of resource namespaces in a ResourceGroup CR",
		stats.UnitDimensionless)

	// ClusterScopedResourceCount tracks the number of cluster-scoped resources in a ResourceGroup CR.
	// This metric should be updated in the Root controller.
	ClusterScopedResourceCount = stats.Int64(
		"cluster_scoped_resource_count",
		"The number of cluster-scoped resources in a ResourceGroup CR",
		stats.UnitDimensionless)

	// CRDCount tracks the number of CRDs in a ResourceGroup CR.
	// This metric should be updated in the Root controller.
	CRDCount = stats.Int64(
		"crd_count",
		"The number of CRDs in a ResourceGroup CR",
		stats.UnitDimensionless)

	// PipelineError tracks the error that happened when syncing a commit
	PipelineError = stats.Int64(
		"pipeline_error_observed",
		"A boolean value indicates if error happened at readiness stage when syncing a commit",
		stats.UnitDimensionless)
)

var (
	// KeyStallReason groups metrics by the stall condition reason field
	KeyStallReason, _ = tag.NewKey("stallreason")

	// KeyOperation groups metrics by their operation. Possible values: create, patch, update, delete.
	KeyOperation, _ = tag.NewKey("operation")

	// KeyErrorCode groups metrics by their error code.
	KeyErrorCode, _ = tag.NewKey("errorcode")

	// KeyType groups metrics by their resource reconciler type. Possible values: root-sync, repo-sync
	KeyType, _ = tag.NewKey("reconciler")

	// KeyResourceGroup groups metrics by their resource group
	KeyResourceGroup, _ = tag.NewKey("resourcegroup")

	// KeyName groups metrics by their name of reconciler.
	KeyName, _ = tag.NewKey("name")

	// KeyComponent groups metrics by their component. Possible value: readiness
	KeyComponent, _ = tag.NewKey("component")

	// ResourceKeyDeploymentName groups metrics by k8s deployment name.
	// This metric tag is populated from the k8s.deployment.name resource
	// attribute for Prometheus using the resource_to_telemetry_conversion feature.
	ResourceKeyDeploymentName, _ = tag.NewKey("k8s_deployment_name")
)

var (
	// ReconcileDurationView aggregates the ReconcileDuration metric measurements.
	ReconcileDurationView = &view.View{
		Name:        ReconcileDuration.Name(),
		Measure:     ReconcileDuration,
		Description: "The distribution of time taken to reconcile a ResourceGroup CR",
		TagKeys:     []tag.Key{KeyStallReason},
		Aggregation: view.Distribution(0.05, 0.1, 0.25, 0.5, 1, 2.5, 5, 10),
	}

	// ResourceGroupTotalView aggregates the ResourceGroupTotal metric measurements.
	ResourceGroupTotalView = &view.View{
		Name:        ResourceGroupTotal.Name(),
		Measure:     ResourceGroupTotal,
		Description: "The current number of ResourceGroup CRs",
		Aggregation: view.LastValue(),
	}

	// ResourceCountView aggregates the ResourceCount metric measurements.
	ResourceCountView = &view.View{
		Name:        ResourceCount.Name(),
		Measure:     ResourceCount,
		Description: "The total number of resources tracked by a ResourceGroup",
		TagKeys:     []tag.Key{KeyResourceGroup},
		Aggregation: view.LastValue(),
	}

	// ReadyResourceCountView aggregates the resources ready in a ResourceGroup
	ReadyResourceCountView = &view.View{
		Name:        ReadyResourceCount.Name(),
		Measure:     ReadyResourceCount,
		Description: "The total number of ready resources in a ResourceGroup",
		TagKeys:     []tag.Key{KeyResourceGroup},
		Aggregation: view.LastValue(),
	}

	// NamespaceCountView counts number of namespaces in a ResourceGroup
	NamespaceCountView = &view.View{
		Name:        NamespaceCount.Name(),
		Measure:     NamespaceCount,
		Description: "The number of namespaces used by resources in a ResourceGroup",
		TagKeys:     []tag.Key{KeyResourceGroup},
		Aggregation: view.LastValue(),
	}

	// ClusterScopedResourceCountView counts number of namespaces in a ResourceGroup
	ClusterScopedResourceCountView = &view.View{
		Name:        ClusterScopedResourceCount.Name(),
		Measure:     ClusterScopedResourceCount,
		Description: "The number of cluster scoped resources in a ResourceGroup",
		TagKeys:     []tag.Key{KeyResourceGroup},
		Aggregation: view.LastValue(),
	}

	// CRDCountView counts number of namespaces in a ResourceGroup
	CRDCountView = &view.View{
		Name:        CRDCount.Name(),
		Measure:     CRDCount,
		Description: "The number of CRDs in a ResourceGroup",
		TagKeys:     []tag.Key{KeyResourceGroup},
		Aggregation: view.LastValue(),
	}

	// KCCResourceCountView aggregates the KCC resources in a ResourceGroup
	KCCResourceCountView = &view.View{
		Name:        KCCResourceCount.Name(),
		Measure:     KCCResourceCount,
		Description: "The total number of KCC resources in a ResourceGroup",
		TagKeys:     []tag.Key{KeyResourceGroup},
		Aggregation: view.LastValue(),
	}

	// PipelineErrorView aggregates the PipelineError by components
	// TODO: add link to same metric in Config Sync under pkg/metrics/views.go
	PipelineErrorView = &view.View{
		Name:        PipelineError.Name(),
		Measure:     PipelineError,
		Description: "A boolean value indicates if error happened from different stages when syncing a commit",
		TagKeys:     []tag.Key{KeyName, KeyComponent, KeyType},
		Aggregation: view.LastValue(),
	}
)

Functions

func ComputeReconcilerNameType

func ComputeReconcilerNameType(nn types.NamespacedName) (reconcilerName, reconcilerType string)

ComputeReconcilerNameType computes the reconciler name from the ResourceGroup CR name

func RecordCRDCount

func RecordCRDCount(ctx context.Context, nn types.NamespacedName, count int64)

RecordCRDCount produces a measurement for RecordCRDCount view

func RecordClusterScopedResourceCount

func RecordClusterScopedResourceCount(ctx context.Context, nn types.NamespacedName, count int64)

RecordClusterScopedResourceCount produces a measurement for ClusterScopedResourceCount view

func RecordKCCResourceCount

func RecordKCCResourceCount(ctx context.Context, nn types.NamespacedName, count int64)

RecordKCCResourceCount produces a measurement for the KCCResourceCount view.

func RecordNamespaceCount

func RecordNamespaceCount(ctx context.Context, nn types.NamespacedName, count int64)

RecordNamespaceCount produces a measurement for the NamespaceCount view.

func RecordPipelineError

func RecordPipelineError(ctx context.Context, nn types.NamespacedName, component string, hasErr bool)

RecordPipelineError produces a measurement for PipelineErrorView

func RecordReadyResourceCount

func RecordReadyResourceCount(ctx context.Context, nn types.NamespacedName, count int64)

RecordReadyResourceCount produces a measurement for the ReadyResourceCount view.

func RecordReconcileDuration

func RecordReconcileDuration(ctx context.Context, stallStatus string, startTime time.Time)

RecordReconcileDuration produces a measurement for the ReconcileDuration view.

func RecordResourceCount

func RecordResourceCount(ctx context.Context, nn types.NamespacedName, count int64)

RecordResourceCount produces a measurement for the ResourceCount view.

func RecordResourceGroupTotal

func RecordResourceGroupTotal(ctx context.Context, count int64)

RecordResourceGroupTotal produces a measurement for the ResourceGroupTotalView

func RegisterOCAgentExporter

func RegisterOCAgentExporter() (*ocagent.Exporter, error)

RegisterOCAgentExporter creates the OC Agent metrics exporter.

func RegisterReconcilerMetricsViews

func RegisterReconcilerMetricsViews() error

RegisterReconcilerMetricsViews registers the views so that recorded metrics can be exported in the reconcilers.