Skip to main content

clock_freq

Overview​

Validates GPU and memory clock frequencies meet minimum application requirements using nvmlDeviceGetClockInfo(). Ensures GPUs operate at expected performance levels for workload execution.

Command-Line Options​

OptionTypeDefaultDescription
--gpu_app_freqInteger1155Minimum GPU clock frequency (MHz)
--gpu_app_mem_freqInteger1593Minimum memory clock frequency (MHz)
--timeoutInteger300Command execution timeout in seconds
--sinkStringdo_nothingTelemetry sink destination
--sink-optsMultiple-Sink-specific configuration
--verbose-outFlagFalseDisplay detailed output
--log-levelChoiceINFODEBUG, INFO, WARNING, ERROR, CRITICAL
--log-folderString/var/log/fb-monitoringLog directory
--heterogeneous-cluster-v1FlagFalseEnable heterogeneous cluster support

Exit Conditions​

Exit CodeCondition
OK (0)Feature flag disabled (killswitch active)
OK (0)All GPUs meet frequency requirements
CRITICAL (2)Any GPU below threshold
UNKNOWN (3)NVML initialization failure

Usage Examples​

clock_freq - Basic Check​

health_checks check-nvidia-smi \
-c clock_freq \
--sink otel \
--sink-opts "log_resource_attributes={'attr_1': 'value1'}" \
[CLUSTER] \
app

clock_freq - Custom Frequencies​

health_checks check-nvidia-smi \
-c clock_freq \
--gpu_app_freq 1410 \
--gpu_app_mem_freq 1800 \
--sink otel \
--sink-opts "log_resource_attributes={'attr_1': 'value1'}" \
[CLUSTER] \
app