metrics

package
v1.16.3 Latest Latest
Warning

This package is not in the latest version of its module.

 Go to latest Published: Oct 23, 2019 License: Apache-2.0 Imports: 31 Imported by: 0

Documentation

Index

Constants

View Source 
const (
	// SingleCallTimeout is how long to try single API calls (like 'get' or 'list'). Used to prevent
	// transient failures from failing tests.
	// TODO: client should not apply this timeout to Watch calls. Increased from 30s until that is fixed.
	SingleCallTimeout = 5 * time.Minute
)

Variables

This section is empty.

Functions

func APICallToPerfData 

func APICallToPerfData(apicalls *APIResponsiveness) *e2eperftype.PerfData

APICallToPerfData transforms APIResponsiveness to PerfData.

func LogSuspiciousLatency 

func LogSuspiciousLatency(latencyData []PodLatencyData, latencyDataLag []PodLatencyData, nodeCount int, c clientset.Interface)

LogSuspiciousLatency logs metrics/docker errors from all nodes that had slow startup times If latencyDataLag is nil then it will be populated from latencyData

func PodStartupLatencyToPerfData 

func PodStartupLatencyToPerfData(latency *PodStartupLatency) *e2eperftype.PerfData

PodStartupLatencyToPerfData transforms PodStartupLatency to PerfData.

func PrettyPrintJSON 

func PrettyPrintJSON(metrics interface{}) string

PrettyPrintJSON converts metrics to JSON format. TODO: This function should be replaced with framework.PrettyPrintJSON after solving circulary dependency between core framework and this metrics subpackage.

func PrintLatencies 

func PrintLatencies(latencies []PodLatencyData, header string)

PrintLatencies outputs latencies to log with readable format.

func ResetMetrics 

func ResetMetrics(c clientset.Interface) error

ResetMetrics resets latency metrics in apiserver.

func ResetSchedulerMetrics 

func ResetSchedulerMetrics(c clientset.Interface, provider, cloudMasterName, masterHostname string) error

ResetSchedulerMetrics sends a DELETE request to kube-scheduler for resetting metrics.

func VerifyLatencyWithinThreshold 

func VerifyLatencyWithinThreshold(threshold, actual LatencyMetric, metricName string) error

VerifyLatencyWithinThreshold verifies whether 50, 90 and 99th percentiles of a latency metric are within the expected threshold.

Types

type APICall 

type APICall struct {
	Resource    string        `json:"resource"`
	Subresource string        `json:"subresource"`
	Verb        string        `json:"verb"`
	Scope       string        `json:"scope"`
	Latency     LatencyMetric `json:"latency"`
	Count       int           `json:"count"`
}

APICall is a struct for managing API call.

type APIResponsiveness 

type APIResponsiveness struct {
	APICalls []APICall `json:"apicalls"`
}

APIResponsiveness is a struct for managing multiple API calls.

func HighLatencyRequests 

func HighLatencyRequests(c clientset.Interface, nodeCount int) (int, *APIResponsiveness, error)

HighLatencyRequests prints top five summary metrics for request types with latency and returns number of such request types above threshold. We use a higher threshold for list calls if nodeCount is above a given threshold (i.e. cluster is big).

func (*APIResponsiveness) Len 

func (a *APIResponsiveness) Len() int

func (*APIResponsiveness) Less 

func (a *APIResponsiveness) Less(i, j int) bool

func (*APIResponsiveness) PrintHumanReadable 

func (a *APIResponsiveness) PrintHumanReadable() string

PrintHumanReadable returns metrics with JSON format.

func (*APIResponsiveness) PrintJSON 

func (a *APIResponsiveness) PrintJSON() string

PrintJSON returns metrics of PerfData(50, 90 and 99th percentiles) with JSON format.

func (*APIResponsiveness) SummaryKind 

func (a *APIResponsiveness) SummaryKind() string

SummaryKind returns the summary of API responsiveness.

func (*APIResponsiveness) Swap 

func (a *APIResponsiveness) Swap(i, j int)

type APIServerMetrics 

type APIServerMetrics Metrics

APIServerMetrics is metrics for API server

func (*APIServerMetrics) Equal 

func (m *APIServerMetrics) Equal(o APIServerMetrics) bool

Equal returns true if all metrics are the same as the arguments.

type ClusterAutoscalerMetrics 

type ClusterAutoscalerMetrics Metrics

ClusterAutoscalerMetrics is metrics for cluster autoscaler

func (*ClusterAutoscalerMetrics) Equal 

func (m *ClusterAutoscalerMetrics) Equal(o ClusterAutoscalerMetrics) bool

Equal returns true if all metrics are the same as the arguments.

type Collection 

type Collection struct {
	APIServerMetrics         APIServerMetrics
	ControllerManagerMetrics ControllerManagerMetrics
	KubeletMetrics           map[string]KubeletMetrics
	SchedulerMetrics         SchedulerMetrics
	ClusterAutoscalerMetrics ClusterAutoscalerMetrics
}

Collection is metrics collection of components

type ComponentCollection 

type ComponentCollection Collection

ComponentCollection is metrics collection of components.

func (*ComponentCollection) ComputeClusterAutoscalerMetricsDelta 

func (m *ComponentCollection) ComputeClusterAutoscalerMetricsDelta(before Collection)

ComputeClusterAutoscalerMetricsDelta computes the change in cluster autoscaler metrics.

func (*ComponentCollection) PrintHumanReadable 

func (m *ComponentCollection) PrintHumanReadable() string

PrintHumanReadable returns e2e metrics with JSON format.

func (*ComponentCollection) PrintJSON 

func (m *ComponentCollection) PrintJSON() string

PrintJSON returns e2e metrics with JSON format.

func (*ComponentCollection) SummaryKind 

func (m *ComponentCollection) SummaryKind() string

SummaryKind returns the summary of e2e metrics.

type ControllerManagerMetrics 

type ControllerManagerMetrics Metrics

ControllerManagerMetrics is metrics for controller manager

func (*ControllerManagerMetrics) Equal 

func (m *ControllerManagerMetrics) Equal(o ControllerManagerMetrics) bool

Equal returns true if all metrics are the same as the arguments.

type EtcdMetrics 

type EtcdMetrics struct {
	BackendCommitDuration     HistogramVec `json:"backendCommitDuration"`
	SnapshotSaveTotalDuration HistogramVec `json:"snapshotSaveTotalDuration"`
	PeerRoundTripTime         HistogramVec `json:"peerRoundTripTime"`
	WalFsyncDuration          HistogramVec `json:"walFsyncDuration"`
	MaxDatabaseSize           float64      `json:"maxDatabaseSize"`
}

EtcdMetrics is a struct for managing etcd metrics.

func (*EtcdMetrics) PrintHumanReadable 

func (l *EtcdMetrics) PrintHumanReadable() string

PrintHumanReadable returns etcd metrics with JSON format.

func (*EtcdMetrics) PrintJSON 

func (l *EtcdMetrics) PrintJSON() string

PrintJSON returns etcd metrics with JSON format.

func (*EtcdMetrics) SummaryKind 

func (l *EtcdMetrics) SummaryKind() string

SummaryKind returns the summary of etcd metrics.

type EtcdMetricsCollector 

type EtcdMetricsCollector struct {
	// contains filtered or unexported fields
}

EtcdMetricsCollector is a struct for managing etcd metrics collector.

func NewEtcdMetricsCollector 

func NewEtcdMetricsCollector() *EtcdMetricsCollector

NewEtcdMetricsCollector creates a new etcd metrics collector.

func (*EtcdMetricsCollector) GetMetrics 

func (mc *EtcdMetricsCollector) GetMetrics() *EtcdMetrics

GetMetrics returns metrics of etcd metrics collector.

func (*EtcdMetricsCollector) StartCollecting 

func (mc *EtcdMetricsCollector) StartCollecting(interval time.Duration, provider string, masterHostname string)

StartCollecting starts to collect etcd db size metric periodically and updates MaxDatabaseSize accordingly.

func (*EtcdMetricsCollector) StopAndSummarize 

func (mc *EtcdMetricsCollector) StopAndSummarize(provider string, masterHostname string) error

StopAndSummarize stops etcd metrics collector and summarizes the metrics.

type Grabber 

type Grabber struct {
	// contains filtered or unexported fields
}

Grabber provides functions which grab metrics from components

func NewMetricsGrabber 

func NewMetricsGrabber(c clientset.Interface, ec clientset.Interface, kubelets bool, scheduler bool, controllers bool, apiServer bool, clusterAutoscaler bool) (*Grabber, error)

NewMetricsGrabber returns new metrics which are initialized.

func (*Grabber) Grab 

func (g *Grabber) Grab() (Collection, error)

Grab returns metrics from corresponding component

func (*Grabber) GrabFromAPIServer 

func (g *Grabber) GrabFromAPIServer() (APIServerMetrics, error)

GrabFromAPIServer returns metrics from API server

func (*Grabber) GrabFromClusterAutoscaler 

func (g *Grabber) GrabFromClusterAutoscaler() (ClusterAutoscalerMetrics, error)

GrabFromClusterAutoscaler returns metrics from cluster autoscaler

func (*Grabber) GrabFromControllerManager 

func (g *Grabber) GrabFromControllerManager() (ControllerManagerMetrics, error)

GrabFromControllerManager returns metrics from controller manager

func (*Grabber) GrabFromKubelet 

func (g *Grabber) GrabFromKubelet(nodeName string) (KubeletMetrics, error)

GrabFromKubelet returns metrics from kubelet

func (*Grabber) GrabFromScheduler 

func (g *Grabber) GrabFromScheduler() (SchedulerMetrics, error)

GrabFromScheduler returns metrics from scheduler

func (*Grabber) HasRegisteredMaster 

func (g *Grabber) HasRegisteredMaster() bool

HasRegisteredMaster returns if metrics grabber was able to find a master node

type Histogram 

type Histogram struct {
	Labels  map[string]string `json:"labels"`
	Buckets map[string]int    `json:"buckets"`
}

Histogram is a struct for managing histogram.

type HistogramVec 

type HistogramVec []Histogram

HistogramVec is an array of Histogram.

type KubeletLatencyMetric 

type KubeletLatencyMetric struct {
	// eg: list, info, create
	Operation string
	// eg: sync_pods, pod_worker
	Method string
	// 0 <= quantile <=1, e.g. 0.95 is 95%tile, 0.5 is median.
	Quantile float64
	Latency  time.Duration
}

KubeletLatencyMetric stores metrics scraped from the kubelet server's /metric endpoint. TODO: Get some more structure around the metrics and this type

type KubeletLatencyMetrics 

type KubeletLatencyMetrics []KubeletLatencyMetric

KubeletLatencyMetrics implements sort.Interface for []KubeletMetric based on the latency field.

func GetDefaultKubeletLatencyMetrics 

func GetDefaultKubeletLatencyMetrics(ms KubeletMetrics) KubeletLatencyMetrics

GetDefaultKubeletLatencyMetrics calls GetKubeletLatencyMetrics with a set of default metricNames identifying common latency metrics. Note that the KubeletMetrics passed in should not contain subsystem prefix.

func GetKubeletLatencyMetrics 

func GetKubeletLatencyMetrics(ms KubeletMetrics, filterMetricNames sets.String) KubeletLatencyMetrics

GetKubeletLatencyMetrics filters ms to include only those contained in the metricNames set, then constructs a KubeletLatencyMetrics list based on the samples associated with those metrics.

func HighLatencyKubeletOperations 

func HighLatencyKubeletOperations(c clientset.Interface, threshold time.Duration, nodeName string, logFunc func(fmt string, args ...interface{})) (KubeletLatencyMetrics, error)

HighLatencyKubeletOperations logs and counts the high latency metrics exported by the kubelet server via /metrics.

func (KubeletLatencyMetrics) Len 

func (a KubeletLatencyMetrics) Len() int

func (KubeletLatencyMetrics) Less 

func (a KubeletLatencyMetrics) Less(i, j int) bool

func (KubeletLatencyMetrics) Swap 

func (a KubeletLatencyMetrics) Swap(i, j int)

type KubeletMetrics 

type KubeletMetrics Metrics

KubeletMetrics is metrics for kubelet

func GetKubeletMetrics 

func GetKubeletMetrics(c clientset.Interface, nodeName string) (KubeletMetrics, error)

GetKubeletMetrics gets all metrics in kubelet subsystem from specified node and trims the subsystem prefix.

func GrabKubeletMetricsWithoutProxy 

func GrabKubeletMetricsWithoutProxy(nodeName, path string) (KubeletMetrics, error)

GrabKubeletMetricsWithoutProxy retrieve metrics from the kubelet on the given node using a simple GET over http. Currently only used in integration tests.

func NewKubeletMetrics 

func NewKubeletMetrics() KubeletMetrics

NewKubeletMetrics returns new metrics which are initialized.

func (*KubeletMetrics) Equal 

func (m *KubeletMetrics) Equal(o KubeletMetrics) bool

Equal returns true if all metrics are the same as the arguments.

type LatencyMetric 

type LatencyMetric struct {
	Perc50  time.Duration `json:"Perc50"`
	Perc90  time.Duration `json:"Perc90"`
	Perc99  time.Duration `json:"Perc99"`
	Perc100 time.Duration `json:"Perc100"`
}

LatencyMetric is a struct for dashboard metrics.

func ExtractLatencyMetrics 

func ExtractLatencyMetrics(latencies []PodLatencyData) LatencyMetric

ExtractLatencyMetrics returns latency metrics for each percentile(50th, 90th and 99th).

type LatencySlice 

type LatencySlice []PodLatencyData

LatencySlice is an array of PodLatencyData which encapsulates pod startup latency information.

func (LatencySlice) Len 

func (a LatencySlice) Len() int

func (LatencySlice) Less 

func (a LatencySlice) Less(i, j int) bool

func (LatencySlice) Swap 

func (a LatencySlice) Swap(i, j int)

type Metrics 

type Metrics map[string]model.Samples

Metrics is generic metrics for other specific metrics

func NewMetrics 

func NewMetrics() Metrics

NewMetrics returns new metrics which are initialized.

func (*Metrics) Equal 

func (m *Metrics) Equal(o Metrics) bool

Equal returns true if all metrics are the same as the arguments.

type PodLatencyData 

type PodLatencyData struct {
	// Name of the pod
	Name string
	// Node this pod was running on
	Node string
	// Latency information related to pod startuptime
	Latency time.Duration
}

PodLatencyData encapsulates pod startup latency information.

type PodStartupLatency 

type PodStartupLatency struct {
	CreateToScheduleLatency LatencyMetric `json:"createToScheduleLatency"`
	ScheduleToRunLatency    LatencyMetric `json:"scheduleToRunLatency"`
	RunToWatchLatency       LatencyMetric `json:"runToWatchLatency"`
	ScheduleToWatchLatency  LatencyMetric `json:"scheduleToWatchLatency"`
	E2ELatency              LatencyMetric `json:"e2eLatency"`
}

PodStartupLatency is a struct for managing latency of pod startup.

func (*PodStartupLatency) PrintHumanReadable 

func (l *PodStartupLatency) PrintHumanReadable() string

PrintHumanReadable returns pod startup letency with JSON format.

func (*PodStartupLatency) PrintJSON 

func (l *PodStartupLatency) PrintJSON() string

PrintJSON returns pod startup letency with JSON format.

func (*PodStartupLatency) SummaryKind 

func (l *PodStartupLatency) SummaryKind() string

SummaryKind returns the summary of pod startup latency.

type SchedulerMetrics 

type SchedulerMetrics Metrics

SchedulerMetrics is metrics for scheduler

func (*SchedulerMetrics) Equal 

func (m *SchedulerMetrics) Equal(o SchedulerMetrics) bool

Equal returns true if all metrics are the same as the arguments.

type SchedulingMetrics 

type SchedulingMetrics struct {
	PredicateEvaluationLatency  LatencyMetric `json:"predicateEvaluationLatency"`
	PriorityEvaluationLatency   LatencyMetric `json:"priorityEvaluationLatency"`
	PreemptionEvaluationLatency LatencyMetric `json:"preemptionEvaluationLatency"`
	BindingLatency              LatencyMetric `json:"bindingLatency"`
	ThroughputAverage           float64       `json:"throughputAverage"`
	ThroughputPerc50            float64       `json:"throughputPerc50"`
	ThroughputPerc90            float64       `json:"throughputPerc90"`
	ThroughputPerc99            float64       `json:"throughputPerc99"`
}

SchedulingMetrics is a struct for managing scheduling metrics.

func VerifySchedulerLatency 

func VerifySchedulerLatency(c clientset.Interface, provider, cloudMasterName, masterHostname string) (*SchedulingMetrics, error)

VerifySchedulerLatency verifies (currently just by logging them) the scheduling latencies.

func (*SchedulingMetrics) PrintHumanReadable 

func (l *SchedulingMetrics) PrintHumanReadable() string

PrintHumanReadable returns scheduling metrics with JSON format.

func (*SchedulingMetrics) PrintJSON 

func (l *SchedulingMetrics) PrintJSON() string

PrintJSON returns scheduling metrics with JSON format.

func (*SchedulingMetrics) SummaryKind 

func (l *SchedulingMetrics) SummaryKind() string

SummaryKind returns the summary of scheduling metrics.

Source Files

View all Source files

Jump to

Keyboard shortcuts

? : This menu
/ : Search site
f or F : Jump to
y or Y : Canonical URL
go.dev uses cookies from Google to deliver and enhance the quality of its services and to analyze traffic. Learn more.