config

Documentation

Index

• Constants
• Variables
• func GetBindAddress(cmdSet string) string
• func GetCGroupVersion() int
• func GetKernelSourceDirs() []string
• func GetMetricPath(cmdSet string) string
• func GetModelConfig(modelItem string) (useEstimatorSidecar bool, selectedModel, selectFilter, initModelURL string)
• func GetRedfishCredFilePath() string
• func GetRedfishProbeIntervalInSeconds() int
• func GetRedfishSkipSSLVerify() bool
• func InitModelConfigMap()
• func LogConfigs()
• func SetEnableAPIServer(enabled bool)
• func SetEnabledEBPFCgroupID(enabled bool)
• func SetEnabledGPU(enabled bool)
• func SetEnabledHardwareCounterMetrics(enabled bool)
• func SetEstimatorConfig(modelName, selectFilter string)
• func SetKernelSourceDir(dir string) error
• func SetKubeConfig(k string)
• func SetModelServerEndpoint(serverEndpoint string)
• func SetModelServerReqEndpoint() (modelServerReqEndpoint string)
• func SetRedfishCredFilePath(credFilePath string)
• func SetRedfishProbeIntervalInSeconds(interval string)
• func SetRedfishSkipSSLVerify(skipSSLVerify bool)
• type Client

Constants

const (

	// MaxIRQ is the maximum number of IRQs to be monitored
	MaxIRQ = 10

	// SamplePeriodSec is the time in seconds that the reader will wait before reading the metrics again
	SamplePeriodSec = 3
)

const (
	// counter - attacher package
	CPUCycle       = "cpu_cycles"
	CPURefCycle    = "cpu_ref_cycles"
	CPUInstruction = "cpu_instr"
	CacheMiss      = "cache_miss"

	// bpf - attacher package
	CPUTime       = "cpu_time"
	IRQNetTXLabel = "irq_net_tx"
	IRQNetRXLabel = "irq_net_rx"
	IRQBlockLabel = "irq_block"

	// cgroup - cgroup package
	CgroupfsMemory       = "cgroupfs_memory_usage_bytes"
	CgroupfsKernelMemory = "cgroupfs_kernel_memory_usage_bytes"
	CgroupfsTCPMemory    = "cgroupfs_tcp_memory_usage_bytes"
	CgroupfsCPU          = "cgroupfs_cpu_usage_us"
	CgroupfsSystemCPU    = "cgroupfs_system_cpu_usage_us"
	CgroupfsUserCPU      = "cgroupfs_user_cpu_usage_us"
	CgroupfsReadIO       = "cgroupfs_ioread_bytes"
	CgroupfsWriteIO      = "cgroupfs_iowrite_bytes"
	BytesReadIO          = "bytes_read"
	BytesWriteIO         = "bytes_writes"
	BlockDevicesIO       = "block_devices_used"
	// kubelet - package
	KubeletContainerCPU    = "container_cpu_usage_seconds_total"
	KubeletContainerMemory = "container_memory_working_set_bytes"
	KubeletNodeCPU         = "node_cpu_usage_seconds_total"
	KubeletNodeMemory      = "node_memory_working_set_bytes"

	// system
	CPUFrequency = "avg_cpu_frequency"

	// GPU
	GPUSMUtilization  = "gpu_sm_util"
	GPUMemUtilization = "gpu_mem_util"
)

Variables

var (
	EnabledMSR            = false
	EnabledBPFBatchDelete = true

	KernelVersion = float32(0)

	KeplerNamespace              = getConfig("KEPLER_NAMESPACE", defaultNamespace)
	UseLibBPFAttacher            = getBoolConfig("LIBBPF_ATTACH", false)
	EnabledEBPFCgroupID          = getBoolConfig("ENABLE_EBPF_CGROUPID", true)
	EnabledGPU                   = getBoolConfig("ENABLE_GPU", false)
	EnableProcessMetrics         = getBoolConfig("ENABLE_PROCESS_METRICS", false)
	ExposeHardwareCounterMetrics = getBoolConfig("EXPOSE_HW_COUNTER_METRICS", true)
	ExposeCgroupMetrics          = getBoolConfig("EXPOSE_CGROUP_METRICS", true)
	ExposeKubeletMetrics         = getBoolConfig("EXPOSE_KUBELET_METRICS", true)
	ExposeIRQCounterMetrics      = getBoolConfig("EXPOSE_IRQ_COUNTER_METRICS", true)
	MetricPathKey                = "METRIC_PATH"
	BindAddressKey               = "BIND_ADDRESS"
	CPUArchOverride              = getConfig("CPU_ARCH_OVERRIDE", "")
	MaxLookupRetry               = getIntConfig("MAX_LOOKUP_RETRY", defaultMaxLookupRetry)

	EstimatorModel        = getConfig("ESTIMATOR_MODEL", defaultMetricValue)         // auto-select
	EstimatorSelectFilter = getConfig("ESTIMATOR_SELECT_FILTER", defaultMetricValue) // no filter
	CoreUsageMetric       = getConfig("CORE_USAGE_METRIC", CPUInstruction)
	DRAMUsageMetric       = getConfig("DRAM_USAGE_METRIC", CacheMiss)
	UncoreUsageMetric     = getConfig("UNCORE_USAGE_METRIC", defaultMetricValue) // no metric (evenly divided)
	GpuUsageMetric        = getConfig("GPU_USAGE_METRIC", GPUSMUtilization)      // no metric (evenly divided)
	GeneralUsageMetric    = getConfig("GENERAL_USAGE_METRIC", CPUInstruction)    // for uncategorized energy; pkg - core - uncore

	////////////////////////////////////
	ModelServerEnable   = getBoolConfig("MODEL_SERVER_ENABLE", false)
	ModelServerEndpoint = SetModelServerReqEndpoint()

	// model_item
	NodeTotalKey           = "NODE_TOTAL"
	NodeComponentsKey      = "NODE_COMPONENTS"
	ContainerTotalKey      = "CONTAINER_TOTAL"
	ContainerComponentsKey = "CONTAINER_COMPONENTS"
	ProcessTotalKey        = "PROCESS_TOTAL"
	ProcessComponentsKey   = "PROCESS_COMPONENTS"

	//  attribute
	EstimatorEnabledKey = "ESTIMATOR"
	InitModelURLKey     = "INIT_URL"
	FixedModelNameKey   = "MODEL"
	ModelFiltersKey     = "FILTERS"

	// KubeConfig is used to start k8s client with the pod running outside the cluster
	KubeConfig      = ""
	EnableAPIServer = false

	DefaultDynCompURL = "/var/lib/kepler/data/ScikitMixed.json"
)

Functions

func GetBindAddress

func GetBindAddress(cmdSet string) string

func GetCGroupVersion

func GetCGroupVersion() int

Get cgroup version, return 1 or 2

func GetKernelSourceDirs

func GetKernelSourceDirs() []string

func GetMetricPath

func GetMetricPath(cmdSet string) string

func GetModelConfig

func GetModelConfig(modelItem string) (useEstimatorSidecar bool, selectedModel, selectFilter, initModelURL string)

func GetRedfishCredFilePath

func GetRedfishCredFilePath() string

func GetRedfishProbeIntervalInSeconds

func GetRedfishProbeIntervalInSeconds() int

func GetRedfishSkipSSLVerify

func GetRedfishSkipSSLVerify() bool

func InitModelConfigMap

func InitModelConfigMap()

InitModelConfigMap initializes map of config from MODEL_CONFIG

func LogConfigs

func LogConfigs()

func SetEnableAPIServer

func SetEnableAPIServer(enabled bool)

SetEnableAPIServer enables Kepler to watch apiserver

func SetEnabledEBPFCgroupID

func SetEnabledEBPFCgroupID(enabled bool)

SetEnabledEBPFCgroupID enables the eBPF code to collect cgroup id if the system has kernel version > 4.18

func SetEnabledGPU

func SetEnabledGPU(enabled bool)

SetEnabledGPU enables the exposure of gpu metrics

func SetEnabledHardwareCounterMetrics

func SetEnabledHardwareCounterMetrics(enabled bool)

SetEnabledHardwareCounterMetrics enables the exposure of hardware counter metrics

func SetEstimatorConfig

func SetEstimatorConfig(modelName, selectFilter string)

func SetKernelSourceDir

func SetKernelSourceDir(dir string) error

SetKernelSourceDir sets the directory for all kernel source. This is used for bcc. Only the top level directory is needed.

func SetKubeConfig

func SetKubeConfig(k string)

SetKubeConfig set kubeconfig file

func SetModelServerEndpoint

func SetModelServerEndpoint(serverEndpoint string)

func SetModelServerReqEndpoint

func SetModelServerReqEndpoint() (modelServerReqEndpoint string)

func SetRedfishCredFilePath

func SetRedfishCredFilePath(credFilePath string)

func SetRedfishProbeIntervalInSeconds

func SetRedfishProbeIntervalInSeconds(interval string)

func SetRedfishSkipSSLVerify

func SetRedfishSkipSSLVerify(skipSSLVerify bool)

Types

type Client

type Client interface {
	// contains filtered or unexported methods
}