Nvidia DGX Server (REST)

Description

This connector monitors hardware for Nvidia DGX Servers.

hardware nvidia

Enterprise Connector

This connector requires the Enterprise edition of MetricsHub.

Supersedes

This connector supersedes:

• RedFish

Target

Typical platform: Nvidia DGX

Operating system: Out-Of-Band

This connector is not available for the local host (it is applicable to remote hosts only).

Prerequisites

Leverages: Nvidia DGX REST API

Technology and protocols: HTTP

Examples

CLI

metricshub HOSTNAME -t oob -c +NvidiaDGXREST --https --http-port 443 -u USERNAME

metricshub.yaml

resourceGroups:
  <RESOURCE_GROUP>:
    resources:
      <HOSTNAME-ID>:
        attributes:
          host.name: <HOSTNAME> # Change with actual host name
          host.type: oob
        connectors: [ +NvidiaDGXREST ] # Optional, to load only this connector
        protocols:
          http:
            https: true
            port: 443 # or probably something else
            username: <USERNAME> # Change with actual credentials
            password: <PASSWORD> # Encrypted using metricshub-encrypt

Connector Activation Criteria

The Nvidia DGX Server (REST) connector will be automatically activated, and its status will be reported as OK if all the below criteria are met:

• The HTTP Request below to the managed host succeeds:
  • get /redfish/v1/Systems
  • Request Header: ${file::authHttpHeader.txt}
  • The response body contains: redfish (regex)
• The HTTP Request below to the managed host succeeds:
  • get /redfish/v1/Systems/DGX
  • Request Header: ${file::authHttpHeader.txt}
  • The response body contains: redfish (regex)

Metrics

Type	Collected Metrics	Specific Attributes
cpu	hw.cpu.speed.limit{limit_type="max"} hw.status{hw.type="cpu", state="degraded|failed|ok"} hw.status{hw.type="cpu", state="present"}	hw.parent.id hw.parent.type id model name vendor
enclosure	hw.enclosure.energy hw.enclosure.power hw.power.limit{hw.type="enclosure", limit_type="high.critical"} hw.status{hw.type="enclosure", state="degraded|failed|ok"} hw.status{hw.type="enclosure", state="present"}	id model name serial_number type vendor
fan	hw.fan.speed hw.fan.speed.limit{limit_type="low.critical"} hw.fan.speed_ratio hw.status{hw.type="fan", state="degraded|failed|ok"} hw.status{hw.type="fan", state="present"}	hw.parent.id hw.parent.type id name
gpu	hw.energy{hw.type="gpu"} hw.gpu.speed hw.gpu.speed.limit{limit_type="high.critical"} hw.gpu.speed.limit{limit_type="high.degraded"} hw.gpu.speed.limit{limit_type="low.degraded"} hw.power.limit{hw.type="gpu", limit_type="high.critical"} hw.power{hw.type="gpu"} hw.status{hw.type="gpu", state="degraded|failed|ok"} hw.status{hw.type="gpu", state="present"}	hw.parent.id hw.parent.type id model name serial_number vendor
memory	hw.memory.limit hw.status{hw.type="memory", state="degraded|failed|ok"} hw.status{hw.type="memory", state="present"}	hw.parent.id hw.parent.type id model name serial_number type vendor
network	hw.network.up hw.status{hw.type="network", state="degraded|failed|ok"} hw.status{hw.type="network", state="present"}	hw.parent.id hw.parent.type id name physical_address
physical_disk	hw.physical_disk.size hw.status{hw.type="physical_disk", state="degraded|failed|ok"} hw.status{hw.type="physical_disk", state="present"}	hw.parent.id hw.parent.type id model name vendor
power_supply	hw.power_supply.limit hw.power_supply.power hw.power_supply.utilization hw.status{hw.type="power_supply", state="degraded|failed|ok"} hw.status{hw.type="power_supply", state="present"}	hw.parent.id hw.parent.type id model name power_supply_type serial_number vendor
temperature	hw.status{hw.type="temperature", state="degraded|failed|ok"} hw.status{hw.type="temperature", state="present"} hw.temperature hw.temperature.limit{limit_type="high.critical"} hw.temperature.limit{limit_type="high.degraded"}	hw.parent.id hw.parent.type id name
voltage	hw.status{hw.type="voltage", state="degraded|failed|ok"} hw.status{hw.type="voltage", state="present"} hw.voltage	hw.parent.id hw.parent.type id name