health

Documentation

Index

• func CurrentHealthyTargets(count float64)
• func HealthyTargetsTotal(target string)
• func ReadyzProtocolRequestTotal(code, target string)
• func UnHealthyTargetsTotal(target string)
• type Listener
• type Metrics
  • func Register(registerFn func(...compbasemetrics.Registerable)) *Metrics
• type Notifier
• type Prober
  • func New(targetProvider TargetProvider, restConfig *rest.Config) (*Prober, error)
  • func (sm *Prober) AddListener(listener Listener)
  • func (sm *Prober) Enqueue()
  • func (sm *Prober) Run(ctx context.Context)
  • func (sm *Prober) Targets() ([]string, []string)
  • func (sm *Prober) WithHealthyProbesThreshold(healthyProbesThreshold int) *Prober
  • func (sm *Prober) WithMetrics(metrics *Metrics) *Prober
  • func (sm *Prober) WithProbeInterval(probeInterval time.Duration) *Prober
  • func (sm *Prober) WithProbeResponseTimeout(probeResponseTimeout time.Duration) *Prober
  • func (sm *Prober) WithUnHealthyProbesThreshold(unhealthyProbesThreshold int) *Prober
• type StaticTargetProvider
  • func (sp StaticTargetProvider) CurrentTargetsList() []string
• type TargetProvider

Functions

func CurrentHealthyTargets

func CurrentHealthyTargets(count float64)

CurrentHealthyTargets keeps track of the current number of healthy targets observed by the health monitor

func HealthyTargetsTotal

func HealthyTargetsTotal(target string)

HealthyTargetsTotal increments the total number of healthy instances observed by the health monitor

func ReadyzProtocolRequestTotal

func ReadyzProtocolRequestTotal(code, target string)

ReadyzProtocolRequestTotal increments the total number of requests issues by the health monitor that violate the "readyz" protocol

the "readyz" protocol defines the following HTTP status code:

HTTP 200 - when the server operates normally
HTTP 500 - when the server is not ready, for example, is undergoing a shutdown

func UnHealthyTargetsTotal

func UnHealthyTargetsTotal(target string)

UnHealthyTargetsTotal increments the total number of unhealthy instances observed by the health monitor

Types

type Listener

type Listener interface {
	// Enqueue should be called when an input may have changed
	Enqueue()
}

Listener is an interface to use to notify interested parties of a change.

type Metrics

type Metrics struct {
	// HealthyTargetsTotal increments the total number of healthy instances observed by the health monitor
	HealthyTargetsTotal func(target string)

	// CurrentHealthyTargets keeps track of the current number of healthy targets observed by the health monitor
	CurrentHealthyTargets func(count float64)

	// UnHealthyTargetsTotal increments the total number of unhealthy instances observed by the health monitor
	UnHealthyTargetsTotal func(target string)

	// ReadyzProtocolRequestTotal increments the total number of requests issues by the health monitor that violate the "readyz" protocol
	//
	// the "readyz" protocol defines the following HTTP status code:
	//   HTTP 200 - when the server operates normally
	//   HTTP 500 - when the server is not ready, for example, is undergoing a shutdown
	ReadyzProtocolRequestTotal func(code, target string)
}

Metrics specifies a set of methods that are used to register various metrics

func Register

func Register(registerFn func(...compbasemetrics.Registerable)) *Metrics

Register is a way to register the health monitor related metrics in the provided store

type Notifier

type Notifier interface {
	// AddListener is adds a listener to be notified of potential input changes
	AddListener(listener Listener)
}

Notifier is a way to add listeners

type Prober

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

func New

func New(targetProvider TargetProvider, restConfig *rest.Config) (*Prober, error)

New creates a health monitor that periodically sends requests to the provided targets to check their health.

The following methods allows you to configure behaviour of the monitor after creation.

WithUnHealthyProbesThreshold - that specifies consecutive failed health checks after which a target is considered unhealthy
                               the default value is: 3

WithHealthyProbesThreshold   - that specifies consecutive successful health checks after which a target is considered healthy
                               the default value is: 5

WithProbeResponseTimeout     - that specifies a time limit for requests made by the HTTP client for the health check
                               the default value is: 1 second

WithProbeInterval            - that specifies a time interval at which health checks are send
                               the default value is: 2 seconds

WithMetrics                  - that specifies a set of methods that are used to register various metrics
                               the default value is: no metrics

Additionally the monitor implements Listener and Notifier interfaces.

The health monitor automatically registers for notification if the target provided also implements the Notifier interface. It is implicit so that the provider can provide a static or a dynamic list of targets.

Interested parties can register a listener for notifications about healthy/unhealthy targets changes via AddListener. TODO: instead of restConfig we could accept transport so that it is reused instead of creating a new connection to targets

reusing the transport has the advantage of using the same connection as other clients

func (*Prober) AddListener

func (sm *Prober) AddListener(listener Listener)

AddListener adds a listener to be notified when the list of healthy targets changes

Note: this method is not thread safe and mustn't be called after calling StartMonitoring() method

func (*Prober) Enqueue

func (sm *Prober) Enqueue()

Enqueue schedules refreshing the target list on the next probeInterval This method is used by the TargetProvider to notify that the list has changed

func (*Prober) Run

func (sm *Prober) Run(ctx context.Context)

Run starts monitoring the provided targets until stop channel is closed This method is blocking and it is meant to be launched in a separate goroutine

func (*Prober) Targets

func (sm *Prober) Targets() ([]string, []string)

Targets returns a list of healthy and unhealthy targets

func (*Prober) WithHealthyProbesThreshold

func (sm *Prober) WithHealthyProbesThreshold(healthyProbesThreshold int) *Prober

WithHealthyProbesThreshold specifies consecutive successful health checks after which a target is considered healthy

func (*Prober) WithMetrics

func (sm *Prober) WithMetrics(metrics *Metrics) *Prober

WithMetrics specifies a set of methods that are used to register various metrics

func (*Prober) WithProbeInterval

func (sm *Prober) WithProbeInterval(probeInterval time.Duration) *Prober

WithProbeInterval specifies a time interval at which health checks are send

func (*Prober) WithProbeResponseTimeout

func (sm *Prober) WithProbeResponseTimeout(probeResponseTimeout time.Duration) *Prober

WithProbeResponseTimeout specifies a time limit for requests made by the HTTP client for the health check

func (*Prober) WithUnHealthyProbesThreshold

func (sm *Prober) WithUnHealthyProbesThreshold(unhealthyProbesThreshold int) *Prober

WithUnHealthyProbesThreshold specifies consecutive failed health checks after which a target is considered unhealthy

type StaticTargetProvider

type StaticTargetProvider []string

StaticTargetProvider implements TargetProvider and provides a static list of targets

func (StaticTargetProvider) CurrentTargetsList

func (sp StaticTargetProvider) CurrentTargetsList() []string

type TargetProvider

type TargetProvider interface {
	// CurrentTargetsList returns a precomputed list of targets
	CurrentTargetsList() []string
}

TargetProviders is an interface to use to get a list of targets to monitor