components/otelopscol/receiver/nvmlreceiver/metadata.yaml

# Copyright 2025 Google LLC
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.

type: nvml
status:
  class: receiver
  stability:
    beta: [metrics]

attributes:
  uuid:
    type: string
    description: GPU universally unique identifier

  gpu_number:
    type: string
    description: GPU index starting at 0.

  model:
    type: string
    description: GPU model

  memory_state:
    type: string
    description: GPU memory used or free.
    enum: [used, free]

  pid:
    type: int
    description: Process ID.

  process:
    type: string
    description: Process name.

  command:
    type: string
    description: Process command.

  command_line:
    type: string
    description: Process command line, 1024 characters maximum.

  owner:
    type: string
    description: Process owner.

metrics:
  nvml.gpu.utilization:
    enabled: true
    description: Fraction of time GPU was not idle since the last sample.
    unit: "1"
    gauge:
      value_type: double
    attributes: [model, gpu_number, uuid]

  nvml.gpu.memory.bytes_used:
    enabled: true
    description: Current number of GPU memory bytes used by state. Summing the values of all states yields the total GPU memory space.
    unit: By
    gauge:
      value_type: int
    attributes: [model, gpu_number, uuid, memory_state]

  nvml.gpu.processes.utilization:
    enabled: true
    description: Fraction of time over the process's life thus far during which one or more kernels was executing on the GPU.
    unit: "1"
    gauge:
      value_type: double
    attributes: [model, gpu_number, uuid, pid, process, command, command_line, owner]

  nvml.gpu.processes.max_bytes_used:
    enabled: true
    description: Maximum total GPU memory in bytes that was ever allocated by the process.
    unit: By
    gauge:
      value_type: int
    attributes: [model, gpu_number, uuid, pid, process, command, command_line, owner]