Clara Holoscan Deploy 0.7.4

18.7. Clara Stop RPC

Requests utilization data for all Clara Platform managed GPUs.

rpc Utilization(ClaraUtilizationRequest) returns (stream ClaraUtilizationResponse);


message ClaraUtilizationRequest { RequestHeader header = 1; bool watch = 2; }

18.8.1. Properties

type: message

Standard RPC request header.

See RequestHeader for details. watch

type: bool

Determines if requested GPU utilization data is for a snapshot or streamed.

  • When true, GPU utilization metrics will be continuously streamed to the caller, until the caller terminates the connection.

  • When false, only a single snapshot of current GPU utilization will be provided.


message ClaraUtilizationResponse { ResponseHeader header = 1; repeated GpuUtilization gpu_metrics = 2; }

Clara Deploy SDK Pipelines service will respond to any utilization request with a ClaraUtilizationResponse message. Requestors can use the response message to determine the result of their request.

18.9.1. Properties header

type: message

Standard RPC response header.

See ResponseHeader for details. gpu_metrics

type: list

Utilization data for all Clara Platform managed GPUs.

See GpuUtilization for details.

© Copyright 2018-2021, NVIDIA Corporation. All rights reserved. Last updated on Feb 1, 2023.