Supported Platforms[1]   Connectors Directory[2]

AMD Radeon (ROCm SMI)

Description

This connector provides hardware information about AMD Radeon GPUs.

community[3] hardware[4]

Source[5]

Target

Typical platform: AMD[6]

Operating system: Linux

Prerequisites

Leverages: ROCm drivers with rocm-smi support.

Technology and protocols: Commands

Examples

CLI

metricshub HOSTNAME -t linux -c +AMDRadeon --ssh -u USER

metricshub.yaml

resourceGroups:
  <RESOURCE_GROUP>:
    resources:
      <HOSTNAME-ID>:
        attributes:
          host.name: <HOSTNAME> # Change with actual host name
          host.type: linux
        connectors: [ +AMDRadeon ] # Optional, to load only this connector
        protocols:
          ssh:
            username: <USERNAME> # Change with actual credentials
            password: <PASSWORD> # Encrypted using metricshub-encrypt

Connector Activation Criteria

The AMD Radeon (ROCm SMI) connector will be automatically activated, and its status will be reported as OK if all the below criteria are met:

  • The command below succeeds on the monitored host
    • Command: rocm-smi
    • Output contains: ROCm System Management Interface (regex)

Metrics

Type Collected Metrics Specific Attributes
enclosure
  • hw.status{hw.type="enclosure", state="present"}
    fan
    • hw.fan.speed_ratio
    • hw.status{hw.type="fan", state="present"}
    • hw.parent.type
    • id
    • name
    • sensor_location
    gpu
    • hw.energy{hw.type="gpu"}
    • hw.gpu.memory.bandwidth
    • hw.gpu.memory.utilization
    • hw.gpu.speed
    • hw.gpu.utilization
    • hw.power.limit{hw.type="gpu"}
    • hw.power{hw.type="gpu"}
    • hw.status{hw.type="gpu", state="present"}
    • hw.parent.type
    • id
    • info
    • model
    • name
    • performance_level
    • serial_number
    • vendor
    temperature
    • hw.status{hw.type="temperature", state="present"}
    • hw.temperature
    • hw.parent.type
    • id
    • name
    • sensor_location
    voltage
    • hw.status{hw.type="voltage", state="present"}
    • hw.voltage
    • hw.parent.type
    • id
    • name
    • sensor_location
    No results.