receivers:
  otlp:
    protocols:
      grpc:
  prometheus/nvidia-dcgm:
    config:
      scrape_configs:
        - job_name: 'nvidia-dcgm'
          scrape_interval: 10s
          static_configs:
            - targets:
                - "172.28.32.12:9400"
                - "172.28.32.13:9400"
 
processors:
  batch:
  resource:
    attributes:
      - key: server.address
        from_attribute: net.host.name
        action: insert
      - key: server.port
        from_attribute: net.host.port
        action: insert
      - key: service.name
        value: nvidia-dcgm
        action: update
      - key: INSTANA_PLUGIN
        value: dcgm
        action: insert
      - key: gpu.uuid
        from_attribute: UUID
        action: insert
      - key: gpu.pci_bus_id
        from_attribute: pci_bus_id
        action: insert
      - key: gpu.device
        from_attribute: device
        action: insert
      - key: gpu.model
        from_attribute: modelName
        action: insert
      - key: gpu.hostname
        from_attribute: Hostname
        action: insert
      - key: gpu.driver_version
        from_attribute: DCGM_FI_DRIVER_VERSION
        action: insert
      - key: gpu.sm_clock_mhz
        from_attribute: DCGM_FI_DEV_SM_CLOCK
        action: insert
      - key: gpu.mem_clock_mhz
        from_attribute: DCGM_FI_DEV_MEM_CLOCK
        action: insert
      - key: gpu.memory_temp_c
        from_attribute: DCGM_FI_DEV_MEMORY_TEMP
        action: insert
      - key: gpu.gpu_temp_c
        from_attribute: DCGM_FI_DEV_GPU_TEMP
        action: insert
      - key: gpu.power_w
        from_attribute: DCGM_FI_DEV_POWER_USAGE
        action: insert
      - key: gpu.energy_mj
        from_attribute: DCGM_FI_DEV_TOTAL_ENERGY_CONSUMPTION
        action: insert
      - key: gpu.pcie_retries
        from_attribute: DCGM_FI_DEV_PCIE_REPLAY_COUNTER
        action: insert
      - key: gpu.utilization_percent
        from_attribute: DCGM_FI_DEV_GPU_UTIL
        action: insert
      - key: gpu.memory_util_percent
        from_attribute: DCGM_FI_DEV_MEM_COPY_UTIL
        action: insert
      - key: gpu.encoder_util_percent
        from_attribute: DCGM_FI_DEV_ENC_UTIL
        action: insert
 
exporters:
  otlp:
    endpoint: "localhost:4317"
    tls:
      insecure: true
 
service:
  pipelines:
    metrics/nvidia-dcgm:
      receivers: [prometheus/nvidia-dcgm]
      processors: [batch, resource]
      exporters: [otlp]