Skip to content

GPU Performance Monitoring

Shield360 uses OpenTelemetry to help you monitor NVIDIA and AMD GPUs for AI applications. Track GPU metrics like utilization, temperature, memory usage, and power consumption during AI training and inference workloads.

GPU monitoring can be implemented in two ways depending on your setup and requirements:

Shield360 SDK

It is useful if you already have an AI application running on GPU that’s instrumented with Shield360.

It extends your existing observability to include GPU metrics alongside your LLM traces.

OpenTelemetry GPU Collector

It is useful for remote GPUs with only LLM models hosted, containerized deployments.

This approach allows you to get GPU metrics without modifying application code.

ParameterTypeDefaultDescription
collect_system_metricsbooleanFalseEnable GPU and system metrics collection
otlp_endpointstringNoneOpenTelemetry OTLP endpoint URL
otlp_headersstringNoneAuthentication headers for OTLP endpoint
service_namestring"unknown_service"Name of your AI application
environmentstringNoneDeployment environment (dev, staging, prod)
VariableDescriptionExample
SHIELD360_COLLECT_SYSTEM_METRICSEnable GPU monitoringtrue
OTEL_EXPORTER_OTLP_ENDPOINTOTLP endpoint URLhttp://127.0.0.1:4318
OTEL_SERVICE_NAMEService name for telemetrymy-gpu-app
OTEL_DEPLOYMENT_ENVIRONMENTDeployment environmentproduction