metrics

Documentation

Index

• Constants
• Variables
• func GetGather() metrics.Gatherer
• func Register(collectors ...metrics.StableCollector)
• func SetActiveConfig(source *corev1.NodeConfigSource) error
• func SetAssignedConfig(source *corev1.NodeConfigSource) error
• func SetConfigError(err bool)
• func SetLastKnownGoodConfig(source *corev1.NodeConfigSource) error
• func SetNodeName(name types.NodeName)
• func SinceInSeconds(start time.Time) float64

Constants

const (
	KubeletSubsystem             = "kubelet"
	NodeNameKey                  = "node_name"
	NodeLabelKey                 = "node"
	PodWorkerDurationKey         = "pod_worker_duration_seconds"
	PodStartDurationKey          = "pod_start_duration_seconds"
	CgroupManagerOperationsKey   = "cgroup_manager_duration_seconds"
	PodWorkerStartDurationKey    = "pod_worker_start_duration_seconds"
	PLEGRelistDurationKey        = "pleg_relist_duration_seconds"
	PLEGDiscardEventsKey         = "pleg_discard_events"
	PLEGRelistIntervalKey        = "pleg_relist_interval_seconds"
	PLEGLastSeenKey              = "pleg_last_seen_seconds"
	EvictionsKey                 = "evictions"
	EvictionStatsAgeKey          = "eviction_stats_age_seconds"
	PreemptionsKey               = "preemptions"
	VolumeStatsCapacityBytesKey  = "volume_stats_capacity_bytes"
	VolumeStatsAvailableBytesKey = "volume_stats_available_bytes"
	VolumeStatsUsedBytesKey      = "volume_stats_used_bytes"
	VolumeStatsInodesKey         = "volume_stats_inodes"
	VolumeStatsInodesFreeKey     = "volume_stats_inodes_free"
	VolumeStatsInodesUsedKey     = "volume_stats_inodes_used"
	RunningPodsKey               = "running_pods"
	RunningContainersKey         = "running_containers"
	// Metrics keys of remote runtime operations
	RuntimeOperationsKey         = "runtime_operations_total"
	RuntimeOperationsDurationKey = "runtime_operations_duration_seconds"
	RuntimeOperationsErrorsKey   = "runtime_operations_errors_total"
	// Metrics keys of device plugin operations
	DevicePluginRegistrationCountKey  = "device_plugin_registration_total"
	DevicePluginAllocationDurationKey = "device_plugin_alloc_duration_seconds"
	// Metrics keys of pod resources operations
	PodResourcesEndpointRequestsTotalKey          = "pod_resources_endpoint_requests_total"
	PodResourcesEndpointRequestsListKey           = "pod_resources_endpoint_requests_list"
	PodResourcesEndpointRequestsGetAllocatableKey = "pod_resources_endpoint_requests_get_allocatable"
	PodResourcesEndpointErrorsListKey             = "pod_resources_endpoint_errors_list"
	PodResourcesEndpointErrorsGetAllocatableKey   = "pod_resources_endpoint_errors_get_allocatable"

	// Metric keys for node config
	AssignedConfigKey             = "node_config_assigned"
	ActiveConfigKey               = "node_config_active"
	LastKnownGoodConfigKey        = "node_config_last_known_good"
	ConfigErrorKey                = "node_config_error"
	ConfigSourceLabelKey          = "node_config_source"
	ConfigSourceLabelValueLocal   = "local"
	ConfigUIDLabelKey             = "node_config_uid"
	ConfigResourceVersionLabelKey = "node_config_resource_version"
	KubeletConfigKeyLabelKey      = "node_config_kubelet_key"

	// Metrics keys for RuntimeClass
	RunPodSandboxDurationKey = "run_podsandbox_duration_seconds"
	RunPodSandboxErrorsKey   = "run_podsandbox_errors_total"

	// Metrics to keep track of total number of Pods and Containers started
	StartedPodsTotalKey             = "started_pods_total"
	StartedPodsErrorsTotalKey       = "started_pods_errors_total"
	StartedContainersTotalKey       = "started_containers_total"
	StartedContainersErrorsTotalKey = "started_containers_errors_total"

	// Metrics to track ephemeral container usage by this kubelet
	ManagedEphemeralContainersKey = "managed_ephemeral_containers"

	// Values used in metric labels
	Container          = "container"
	InitContainer      = "init_container"
	EphemeralContainer = "ephemeral_container"
)

This const block defines the metric names for the kubelet metrics.

Variables

var (
	// NodeName is a Gauge that tracks the ode's name. The count is always 1.
	NodeName = metrics.NewGaugeVec(
		&metrics.GaugeOpts{
			Subsystem:      KubeletSubsystem,
			Name:           NodeNameKey,
			Help:           "The node's name. The count is always 1.",
			StabilityLevel: metrics.ALPHA,
		},
		[]string{NodeLabelKey},
	)
	// ContainersPerPodCount is a Histogram that tracks the number of containers per pod.
	ContainersPerPodCount = metrics.NewHistogram(
		&metrics.HistogramOpts{
			Subsystem:      KubeletSubsystem,
			Name:           "containers_per_pod_count",
			Help:           "The number of containers per pod.",
			Buckets:        metrics.ExponentialBuckets(1, 2, 5),
			StabilityLevel: metrics.ALPHA,
		},
	)
	// PodWorkerDuration is a Histogram that tracks the duration (in seconds) in takes to sync a single pod.
	// Broken down by the operation type.
	PodWorkerDuration = metrics.NewHistogramVec(
		&metrics.HistogramOpts{
			Subsystem:      KubeletSubsystem,
			Name:           PodWorkerDurationKey,
			Help:           "Duration in seconds to sync a single pod. Broken down by operation type: create, update, or sync",
			Buckets:        metrics.DefBuckets,
			StabilityLevel: metrics.ALPHA,
		},
		[]string{"operation_type"},
	)
	// PodStartDuration is a Histogram that tracks the duration (in seconds) it takes for a single pod to go from pending to running.
	PodStartDuration = metrics.NewHistogram(
		&metrics.HistogramOpts{
			Subsystem:      KubeletSubsystem,
			Name:           PodStartDurationKey,
			Help:           "Duration in seconds for a single pod to go from pending to running.",
			Buckets:        metrics.DefBuckets,
			StabilityLevel: metrics.ALPHA,
		},
	)
	// CgroupManagerDuration is a Histogram that tracks the duration (in seconds) it takes for cgroup manager operations to complete.
	// Broken down by method.
	CgroupManagerDuration = metrics.NewHistogramVec(
		&metrics.HistogramOpts{
			Subsystem:      KubeletSubsystem,
			Name:           CgroupManagerOperationsKey,
			Help:           "Duration in seconds for cgroup manager operations. Broken down by method.",
			Buckets:        metrics.DefBuckets,
			StabilityLevel: metrics.ALPHA,
		},
		[]string{"operation_type"},
	)
	// PodWorkerStartDuration is a Histogram that tracks the duration (in seconds) it takes from seeing a pod to starting a worker.
	PodWorkerStartDuration = metrics.NewHistogram(
		&metrics.HistogramOpts{
			Subsystem:      KubeletSubsystem,
			Name:           PodWorkerStartDurationKey,
			Help:           "Duration in seconds from seeing a pod to starting a worker.",
			Buckets:        metrics.DefBuckets,
			StabilityLevel: metrics.ALPHA,
		},
	)
	// PLEGRelistDuration is a Histogram that tracks the duration (in seconds) it takes for relisting pods in the Kubelet's
	// Pod Lifecycle Event Generator (PLEG).
	PLEGRelistDuration = metrics.NewHistogram(
		&metrics.HistogramOpts{
			Subsystem:      KubeletSubsystem,
			Name:           PLEGRelistDurationKey,
			Help:           "Duration in seconds for relisting pods in PLEG.",
			Buckets:        metrics.DefBuckets,
			StabilityLevel: metrics.ALPHA,
		},
	)
	// PLEGDiscardEvents is a Counter that tracks the number of discarding events in the Kubelet's Pod Lifecycle Event Generator (PLEG).
	PLEGDiscardEvents = metrics.NewCounter(
		&metrics.CounterOpts{
			Subsystem:      KubeletSubsystem,
			Name:           PLEGDiscardEventsKey,
			Help:           "The number of discard events in PLEG.",
			StabilityLevel: metrics.ALPHA,
		},
	)

	// PLEGRelistInterval is a Histogram that tracks the intervals (in seconds) between relisting in the Kubelet's
	// Pod Lifecycle Event Generator (PLEG).
	PLEGRelistInterval = metrics.NewHistogram(
		&metrics.HistogramOpts{
			Subsystem:      KubeletSubsystem,
			Name:           PLEGRelistIntervalKey,
			Help:           "Interval in seconds between relisting in PLEG.",
			Buckets:        metrics.DefBuckets,
			StabilityLevel: metrics.ALPHA,
		},
	)
	// PLEGLastSeen is a Gauge giving the Unix timestamp when the Kubelet's
	// Pod Lifecycle Event Generator (PLEG) was last seen active.
	PLEGLastSeen = metrics.NewGauge(
		&metrics.GaugeOpts{
			Subsystem:      KubeletSubsystem,
			Name:           PLEGLastSeenKey,
			Help:           "Timestamp in seconds when PLEG was last seen active.",
			StabilityLevel: metrics.ALPHA,
		},
	)
	// RuntimeOperations is a Counter that tracks the cumulative number of remote runtime operations.
	// Broken down by operation type.
	RuntimeOperations = metrics.NewCounterVec(
		&metrics.CounterOpts{
			Subsystem:      KubeletSubsystem,
			Name:           RuntimeOperationsKey,
			Help:           "Cumulative number of runtime operations by operation type.",
			StabilityLevel: metrics.ALPHA,
		},
		[]string{"operation_type"},
	)
	// RuntimeOperationsDuration is a Histogram that tracks the duration (in seconds) for remote runtime operations to complete.
	// Broken down by operation type.
	RuntimeOperationsDuration = metrics.NewHistogramVec(
		&metrics.HistogramOpts{
			Subsystem:      KubeletSubsystem,
			Name:           RuntimeOperationsDurationKey,
			Help:           "Duration in seconds of runtime operations. Broken down by operation type.",
			Buckets:        metrics.ExponentialBuckets(.005, 2.5, 14),
			StabilityLevel: metrics.ALPHA,
		},
		[]string{"operation_type"},
	)
	// RuntimeOperationsErrors is a Counter that tracks the cumulative number of remote runtime operations errors.
	// Broken down by operation type.
	RuntimeOperationsErrors = metrics.NewCounterVec(
		&metrics.CounterOpts{
			Subsystem:      KubeletSubsystem,
			Name:           RuntimeOperationsErrorsKey,
			Help:           "Cumulative number of runtime operation errors by operation type.",
			StabilityLevel: metrics.ALPHA,
		},
		[]string{"operation_type"},
	)
	// Evictions is a Counter that tracks the cumulative number of pod evictions initiated by the kubelet.
	// Broken down by eviction signal.
	Evictions = metrics.NewCounterVec(
		&metrics.CounterOpts{
			Subsystem:      KubeletSubsystem,
			Name:           EvictionsKey,
			Help:           "Cumulative number of pod evictions by eviction signal",
			StabilityLevel: metrics.ALPHA,
		},
		[]string{"eviction_signal"},
	)
	// EvictionStatsAge is a Histogram that tracks the time (in seconds) between when stats are collected and when a pod is evicted
	// based on those stats. Broken down by eviction signal.
	EvictionStatsAge = metrics.NewHistogramVec(
		&metrics.HistogramOpts{
			Subsystem:      KubeletSubsystem,
			Name:           EvictionStatsAgeKey,
			Help:           "Time between when stats are collected, and when pod is evicted based on those stats by eviction signal",
			Buckets:        metrics.DefBuckets,
			StabilityLevel: metrics.ALPHA,
		},
		[]string{"eviction_signal"},
	)
	// Preemptions is a Counter that tracks the cumulative number of pod preemptions initiated by the kubelet.
	// Broken down by preemption signal. A preemption is only recorded for one resource, the sum of all signals
	// is the number of preemptions on the given node.
	Preemptions = metrics.NewCounterVec(
		&metrics.CounterOpts{
			Subsystem:      KubeletSubsystem,
			Name:           PreemptionsKey,
			Help:           "Cumulative number of pod preemptions by preemption resource",
			StabilityLevel: metrics.ALPHA,
		},
		[]string{"preemption_signal"},
	)
	// DevicePluginRegistrationCount is a Counter that tracks the cumulative number of device plugin registrations.
	// Broken down by resource name.
	DevicePluginRegistrationCount = metrics.NewCounterVec(
		&metrics.CounterOpts{
			Subsystem:      KubeletSubsystem,
			Name:           DevicePluginRegistrationCountKey,
			Help:           "Cumulative number of device plugin registrations. Broken down by resource name.",
			StabilityLevel: metrics.ALPHA,
		},
		[]string{"resource_name"},
	)
	// DevicePluginAllocationDuration is a Histogram that tracks the duration (in seconds) to serve a device plugin allocation request.
	// Broken down by resource name.
	DevicePluginAllocationDuration = metrics.NewHistogramVec(
		&metrics.HistogramOpts{
			Subsystem:      KubeletSubsystem,
			Name:           DevicePluginAllocationDurationKey,
			Help:           "Duration in seconds to serve a device plugin Allocation request. Broken down by resource name.",
			Buckets:        metrics.DefBuckets,
			StabilityLevel: metrics.ALPHA,
		},
		[]string{"resource_name"},
	)

	// PodResourcesEndpointRequestsTotalCount is a Counter that tracks the cumulative number of requests to the PodResource endpoints.
	// Broken down by server API version.
	PodResourcesEndpointRequestsTotalCount = metrics.NewCounterVec(
		&metrics.CounterOpts{
			Subsystem:      KubeletSubsystem,
			Name:           PodResourcesEndpointRequestsTotalKey,
			Help:           "Cumulative number of requests to the PodResource endpoint. Broken down by server api version.",
			StabilityLevel: metrics.ALPHA,
		},
		[]string{"server_api_version"},
	)

	// PodResourcesEndpointRequestsListCount is a Counter that tracks the number of requests to the PodResource List() endpoint.
	// Broken down by server API version.
	PodResourcesEndpointRequestsListCount = metrics.NewCounterVec(
		&metrics.CounterOpts{
			Subsystem:      KubeletSubsystem,
			Name:           PodResourcesEndpointRequestsListKey,
			Help:           "Number of requests to the PodResource List endpoint. Broken down by server api version.",
			StabilityLevel: metrics.ALPHA,
		},
		[]string{"server_api_version"},
	)

	// PodResourcesEndpointRequestsGetAllocatableCount is a Counter that tracks the number of requests to the PodResource GetAllocatableResources() endpoint.
	// Broken down by server API version.
	PodResourcesEndpointRequestsGetAllocatableCount = metrics.NewCounterVec(
		&metrics.CounterOpts{
			Subsystem:      KubeletSubsystem,
			Name:           PodResourcesEndpointRequestsGetAllocatableKey,
			Help:           "Number of requests to the PodResource GetAllocatableResources endpoint. Broken down by server api version.",
			StabilityLevel: metrics.ALPHA,
		},
		[]string{"server_api_version"},
	)

	// PodResourcesEndpointErrorsListCount is a Counter that tracks the number of errors returned by he PodResource List() endpoint.
	// Broken down by server API version.
	PodResourcesEndpointErrorsListCount = metrics.NewCounterVec(
		&metrics.CounterOpts{
			Subsystem:      KubeletSubsystem,
			Name:           PodResourcesEndpointErrorsListKey,
			Help:           "Number of requests to the PodResource List endpoint which returned error. Broken down by server api version.",
			StabilityLevel: metrics.ALPHA,
		},
		[]string{"server_api_version"},
	)

	// PodResourcesEndpointErrorsGetAllocatableCount is a Counter that tracks the number of errors returned by the PodResource GetAllocatableResources() endpoint.
	// Broken down by server API version.
	PodResourcesEndpointErrorsGetAllocatableCount = metrics.NewCounterVec(
		&metrics.CounterOpts{
			Subsystem:      KubeletSubsystem,
			Name:           PodResourcesEndpointErrorsGetAllocatableKey,
			Help:           "Number of requests to the PodResource GetAllocatableResources endpoint which returned error. Broken down by server api version.",
			StabilityLevel: metrics.ALPHA,
		},
		[]string{"server_api_version"},
	)

	// AssignedConfig is a Gauge that is set 1 if the Kubelet has a NodeConfig assigned.
	AssignedConfig = metrics.NewGaugeVec(
		&metrics.GaugeOpts{
			Subsystem:         KubeletSubsystem,
			Name:              AssignedConfigKey,
			Help:              "The node's understanding of intended config. The count is always 1.",
			DeprecatedVersion: "1.23.0",
			StabilityLevel:    metrics.ALPHA,
		},
		[]string{ConfigSourceLabelKey, ConfigUIDLabelKey, ConfigResourceVersionLabelKey, KubeletConfigKeyLabelKey},
	)
	// ActiveConfig is a Gauge that is set to 1 if the Kubelet has an active NodeConfig.
	ActiveConfig = metrics.NewGaugeVec(
		&metrics.GaugeOpts{
			Subsystem:         KubeletSubsystem,
			Name:              ActiveConfigKey,
			Help:              "The config source the node is actively using. The count is always 1.",
			DeprecatedVersion: "1.23.0",
			StabilityLevel:    metrics.ALPHA,
		},
		[]string{ConfigSourceLabelKey, ConfigUIDLabelKey, ConfigResourceVersionLabelKey, KubeletConfigKeyLabelKey},
	)
	// LastKnownGoodConfig is a Gauge that is set to 1 if the Kubelet has a NodeConfig it can fall back to if there
	// are certain errors.
	LastKnownGoodConfig = metrics.NewGaugeVec(
		&metrics.GaugeOpts{
			Subsystem:         KubeletSubsystem,
			Name:              LastKnownGoodConfigKey,
			Help:              "The config source the node will fall back to when it encounters certain errors. The count is always 1.",
			DeprecatedVersion: "1.23.0",
			StabilityLevel:    metrics.ALPHA,
		},
		[]string{ConfigSourceLabelKey, ConfigUIDLabelKey, ConfigResourceVersionLabelKey, KubeletConfigKeyLabelKey},
	)
	// ConfigError is a Gauge that is set to 1 if the node is experiencing a configuration-related error.
	ConfigError = metrics.NewGauge(
		&metrics.GaugeOpts{
			Subsystem:         KubeletSubsystem,
			Name:              ConfigErrorKey,
			Help:              "This metric is true (1) if the node is experiencing a configuration-related error, false (0) otherwise.",
			DeprecatedVersion: "1.23.0",
			StabilityLevel:    metrics.ALPHA,
		},
	)
	// RunPodSandboxDuration is a Histogram that tracks the duration (in seconds) it takes to run Pod Sandbox operations.
	// Broken down by RuntimeClass.Handler.
	RunPodSandboxDuration = metrics.NewHistogramVec(
		&metrics.HistogramOpts{
			Subsystem: KubeletSubsystem,
			Name:      RunPodSandboxDurationKey,
			Help:      "Duration in seconds of the run_podsandbox operations. Broken down by RuntimeClass.Handler.",

			Buckets:        metrics.DefBuckets,
			StabilityLevel: metrics.ALPHA,
		},
		[]string{"runtime_handler"},
	)
	// RunPodSandboxErrors is a Counter that tracks the cumulative number of Pod Sandbox operations errors.
	// Broken down by RuntimeClass.Handler.
	RunPodSandboxErrors = metrics.NewCounterVec(
		&metrics.CounterOpts{
			Subsystem:      KubeletSubsystem,
			Name:           RunPodSandboxErrorsKey,
			Help:           "Cumulative number of the run_podsandbox operation errors by RuntimeClass.Handler.",
			StabilityLevel: metrics.ALPHA,
		},
		[]string{"runtime_handler"},
	)

	// RunningPodCount is a gauge that tracks the number of Pods currently with a running sandbox
	// It is used to expose the kubelet internal state: how many pods have running containers in the container runtime, and mainly for debugging purpose.
	RunningPodCount = metrics.NewGauge(
		&metrics.GaugeOpts{
			Subsystem:      KubeletSubsystem,
			Name:           RunningPodsKey,
			Help:           "Number of pods that have a running pod sandbox",
			StabilityLevel: metrics.ALPHA,
		},
	)
	// RunningContainerCount is a gauge that tracks the number of containers currently running
	RunningContainerCount = metrics.NewGaugeVec(
		&metrics.GaugeOpts{
			Subsystem:      KubeletSubsystem,
			Name:           RunningContainersKey,
			Help:           "Number of containers currently running",
			StabilityLevel: metrics.ALPHA,
		},
		[]string{"container_state"},
	)
	// StartedPodsTotal is a counter that tracks pod sandbox creation operations
	StartedPodsTotal = metrics.NewCounter(
		&metrics.CounterOpts{
			Subsystem:      KubeletSubsystem,
			Name:           StartedPodsTotalKey,
			Help:           "Cumulative number of pods started",
			StabilityLevel: metrics.ALPHA,
		},
	)
	// StartedPodsErrorsTotal is a counter that tracks the number of errors creating pod sandboxes
	StartedPodsErrorsTotal = metrics.NewCounterVec(
		&metrics.CounterOpts{
			Subsystem:      KubeletSubsystem,
			Name:           StartedPodsErrorsTotalKey,
			Help:           "Cumulative number of errors when starting pods",
			StabilityLevel: metrics.ALPHA,
		},
		[]string{"message"},
	)
	// StartedContainersTotal is a counter that tracks the number of container creation operations
	StartedContainersTotal = metrics.NewCounterVec(
		&metrics.CounterOpts{
			Subsystem:      KubeletSubsystem,
			Name:           StartedContainersTotalKey,
			Help:           "Cumulative number of containers started",
			StabilityLevel: metrics.ALPHA,
		},
		[]string{"container_type"},
	)
	// StartedContainersTotal is a counter that tracks the number of errors creating containers
	StartedContainersErrorsTotal = metrics.NewCounterVec(
		&metrics.CounterOpts{
			Subsystem:      KubeletSubsystem,
			Name:           StartedContainersErrorsTotalKey,
			Help:           "Cumulative number of errors when starting containers",
			StabilityLevel: metrics.ALPHA,
		},
		[]string{"container_type", "code"},
	)
	// ManagedEphemeralContainers is a gauge that indicates how many ephemeral containers are managed by this kubelet.
	ManagedEphemeralContainers = metrics.NewGauge(
		&metrics.GaugeOpts{
			Subsystem:      KubeletSubsystem,
			Name:           ManagedEphemeralContainersKey,
			Help:           "Current number of ephemeral containers in pods managed by this kubelet. Ephemeral containers will be ignored if disabled by the EphemeralContainers feature gate, and this number will be 0.",
			StabilityLevel: metrics.ALPHA,
		},
	)
)

Functions

func GetGather

func GetGather() metrics.Gatherer

GetGather returns the gatherer. It used by test case outside current package.

func Register

func Register(collectors ...metrics.StableCollector)

Register registers all metrics.

func SetActiveConfig

func SetActiveConfig(source *corev1.NodeConfigSource) error

SetActiveConfig tracks labels according to the NodeConfig that is currently used by the Kubelet. It also tracks labels across metric updates so old labels can be safely deleted.

func SetAssignedConfig

func SetAssignedConfig(source *corev1.NodeConfigSource) error

SetAssignedConfig tracks labels according to the assigned NodeConfig. It also tracks labels across metric updates so old labels can be safely deleted.

func SetConfigError

func SetConfigError(err bool)

SetConfigError sets a the ConfigError metric to 1 in case any errors were encountered.

func SetLastKnownGoodConfig

func SetLastKnownGoodConfig(source *corev1.NodeConfigSource) error

SetLastKnownGoodConfig tracks labels according to the NodeConfig that was successfully applied last. It also tracks labels across metric updates so old labels can be safely deleted.

func SetNodeName

func SetNodeName(name types.NodeName)

SetNodeName sets the NodeName Gauge to 1.

func SinceInSeconds

func SinceInSeconds(start time.Time) float64

SinceInSeconds gets the time since the specified start in seconds.