sidecar

Documentation

Index

• Constants
• type ComponentPowerResponse
• type EstimatorSidecar
  • func (c *EstimatorSidecar) AddContainerFeatureValues(x []float64)
  • func (c *EstimatorSidecar) AddDesiredOutValue(y float64)
  • func (c *EstimatorSidecar) AddNodeFeatureValues(x []float64)
  • func (c *EstimatorSidecar) GetComponentsPower(isIdlePower bool) ([]source.NodeComponentsEnergy, error)
  • func (c *EstimatorSidecar) GetContainerFeatureNamesList() []string
  • func (c *EstimatorSidecar) GetGPUPower(isIdlePower bool) ([]float64, error)
  • func (c *EstimatorSidecar) GetModelType() types.ModelType
  • func (c *EstimatorSidecar) GetNodeFeatureNamesList() []string
  • func (c *EstimatorSidecar) GetPlatformPower(isIdlePower bool) ([]float64, error)
  • func (c *EstimatorSidecar) IsEnabled() bool
  • func (c *EstimatorSidecar) ResetSampleIdx()
  • func (c *EstimatorSidecar) Start() error
  • func (c *EstimatorSidecar) Train() error
• type PlatformPowerResponse
• type PowerRequest

Constants

const MaxContainers = 500 // 256 pods and 2 containers per pod

Types

type ComponentPowerResponse

type ComponentPowerResponse struct {
	Powers  map[string][]float64 `json:"powers"`
	Message string               `json:"msg"`
}

ComponentPowerResponse defines a response of a map of component powers from Kepler Estimator

type EstimatorSidecar

type EstimatorSidecar struct {
	Socket       string
	OutputType   types.ModelOutputType
	EnergySource string
	TrainerName  string
	SelectFilter string

	FloatFeatureNames           []string
	SystemMetaDataFeatureNames  []string
	SystemMetaDataFeatureValues []string
	// contains filtered or unexported fields
}

EstimatorSidecar defines power estimator with Kepler Estimator sidecar

func (*EstimatorSidecar) AddContainerFeatureValues

func (c *EstimatorSidecar) AddContainerFeatureValues(x []float64)

AddContainerFeatureValues adds the the x for prediction, which are the explanatory variables (or the independent variable) of regression. EstimatorSidecar is trained off-line then we cannot Add training samples. We might implement it in the future. The EstimatorSidecar does not differentiate node or container power estimation, the difference will only be the amount of resource utilization

func (*EstimatorSidecar) AddDesiredOutValue

func (c *EstimatorSidecar) AddDesiredOutValue(y float64)

AddDesiredOutValue adds the the y, which is the response variable (or the dependent variable) of regression. EstimatorSidecar is trained off-line then we do not add Y for trainning. We might implement it in the future.

func (*EstimatorSidecar) AddNodeFeatureValues

func (c *EstimatorSidecar) AddNodeFeatureValues(x []float64)

AddNodeFeatureValues adds the the x for prediction, which is the variable used to calculate the ratio. EstimatorSidecar is not trained, then we cannot Add training samples, only samples for prediction. The EstimatorSidecar does not differentiate node or container power estimation, the difference will only be the amount of resource utilization

func (*EstimatorSidecar) GetComponentsPower

func (c *EstimatorSidecar) GetComponentsPower(isIdlePower bool) ([]source.NodeComponentsEnergy, error)

GetComponentsPower makes a request to Kepler Estimator EstimatorSidecar and return a list of total powers

func (*EstimatorSidecar) GetContainerFeatureNamesList

func (c *EstimatorSidecar) GetContainerFeatureNamesList() []string

GetContainerFeatureNamesList returns the list of float features that the model was configured to use The EstimatorSidecar does not differentiate node or container power estimation, the difference will only be the amount of resource utilization

func (*EstimatorSidecar) GetGPUPower

func (c *EstimatorSidecar) GetGPUPower(isIdlePower bool) ([]float64, error)

GetComponentsPower returns GPU Power in Watts associated to each each process/container/pod

func (*EstimatorSidecar) GetModelType

func (c *EstimatorSidecar) GetModelType() types.ModelType

GetModelType returns the model type

func (*EstimatorSidecar) GetNodeFeatureNamesList

func (c *EstimatorSidecar) GetNodeFeatureNamesList() []string

GetNodeFeatureNamesList returns the list of float features that the model was configured to use The EstimatorSidecar does not differentiate node or container power estimation, the difference will only be the amount of resource utilization

func (*EstimatorSidecar) GetPlatformPower

func (c *EstimatorSidecar) GetPlatformPower(isIdlePower bool) ([]float64, error)

GetPlatformPower makes a request to Kepler Estimator EstimatorSidecar and returns a list of total powers

func (*EstimatorSidecar) IsEnabled

func (c *EstimatorSidecar) IsEnabled() bool

IsEnabled returns true if the power model was trained and is active

func (*EstimatorSidecar) ResetSampleIdx

func (c *EstimatorSidecar) ResetSampleIdx()

ResetSampleIdx set the sample vector index to 0 to overwrite the old samples with new ones for trainning or prediction.

func (*EstimatorSidecar) Start

func (c *EstimatorSidecar) Start() error

Start returns nil if estimator is connected and has compatible power model

func (*EstimatorSidecar) Train

func (c *EstimatorSidecar) Train() error

Train triggers the regressiong fit after adding data points to create a new power model. EstimatorSidecar is trained in the Model Server then we cannot trigger the trainning. We might implement it in the future.

type PlatformPowerResponse

type PlatformPowerResponse struct {
	Powers  []float64 `json:"powers"`
	Message string    `json:"msg"`
}

PlatformPowerResponse defines a response of a list of total powers from Kepler Estimator

type PowerRequest

type PowerRequest struct {
	FloatFeatureNames           []string    `json:"metrics"`
	UsageValues                 [][]float64 `json:"values"`
	OutputType                  string      `json:"output_type"`
	EnergySource                string      `json:"source"`
	SystemMetaDataFeatureNames  []string    `json:"system_features"`
	SystemMetaDataFeatureValues []string    `json:"system_values"`
	TrainerName                 string      `json:"trainer_name"`
	SelectFilter                string      `json:"filter"`
}

PowerRequest defines a request to Kepler Estimator to get estimated powers