Get cgroup version, return 1 or 2.

return local path to power model weight e.g., /var/lib/kepler/data/acpi_AbsPowerModel.json.

InitModelConfigMap initializes map of config from MODEL_CONFIG.

IsExposeBPFMetricsEnabled returns false if BPF Metrics metrics are disabled to minimize overhead.

IsExposeComponentPowerEnabled returns false if component power metrics are disabled to minimize overhead.

IsExposeContainerStatsEnabled returns false if container metrics are disabled to minimize overhead in the Kepler standalone mode.

IsExposeProcessStatsEnabled returns false if process metrics are disabled to minimize overhead in the Kepler standalone mode.

IsExposeQATMetricsEnabled returns false if QATMetrics metrics are disabled to minimize overhead.

IsExposeVMStatsEnabled returns false if VM metrics are disabled to minimize overhead.

IsIdlePowerEnabled always return true if Kepler has access to system power metrics.

SetEnableAPIServer enables Kepler to watch apiserver.

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

SetEnabledGPU enables the exposure of gpu metrics.

SetEnabledHardwareCounterMetrics enables the exposure of hardware counter metrics.

SetEnabledIdlePower allows enabling idle power exposure in Kepler's metrics.

SetEnabledQAT enables the exposure of qat metrics.

SetKernelSourceDir sets the directory for all kernel source.

SetKubeConfig set kubeconfig file.

Energy Metrics Absolute energy and power.

counter - attacher package.

bpf - attacher package.

Dynamic energy and power.

NVIDIA GPU.

Idle energy and power.

MaxIRQ is the maximum number of IRQs to be monitored.

Intel QuickAssist Technology (QAT) TODO: test if different request has different energy consumption.

nvidia dcgm hostengine endpoint.

auto-select.

no filter.

for uncategorized energy.

no metric (evenly divided).

KubeConfig is used to start k8s client with the pod running outside the cluster.

for model config.

.

model_parameter_prefix.

the default container power model is RATIO but ESTIMATOR or LINEAR_REGRESSION can be used.

no metric (evenly divided).