Overview

This node collects CPU, GPU, Memory utilization and FPS statistics to profile machine performance, and adds them to metadata, where they can then be exported using Webhooks, Elasticsearch or saved with captured Clips and Snapshots.

Note that this is machine-level performance, so numbers are a cumulative of the number of pipelines and other processes you have running on your machine.

Inputs & Outputs

Inputs : 1, Media Format : Raw Video
Outputs : 1, Media Format: Raw Video
Output Metadata : Machine Performance Statistics

Properties

Property	Description	Type	Default	Required
`interval`	Stats collection interval. Unit: frames.	`number`	`20`	Yes

Output Metadata

The fields below are declared by this node's metadata schema; the JSON values are representative examples.

Path	Type	Description
`nodes.<node_id>.stats.present`	`boolean`	Boolean indicating whether a fresh performance sample is present on this frame.
`nodes.<node_id>.stats.fps`	`integer`	Measured pipeline frame rate in frames per second.
`nodes.<node_id>.stats.process.running_pipelines_count`	`integer`	Number of pipelines currently running on the device.
`nodes.<node_id>.stats.system.is_jetson`	`boolean`	Boolean indicating whether the device is an NVIDIA Jetson platform.
`nodes.<node_id>.stats.system.lumeo_version`	`string`	Installed Lumeo software version.
`nodes.<node_id>.stats.system.cpu_count`	`integer`	Number of logical CPU cores.
`nodes.<node_id>.stats.system.cpu_freq`	`number`	Current CPU frequency in MHz.
`nodes.<node_id>.stats.system.mem_total`	`integer`	Total system memory in GB.
`nodes.<node_id>.stats.system.gpu_count`	`integer`	Number of detected GPUs.
`nodes.<node_id>.stats.system.board.gpus[].type`	`string`	Detected GPU model name.
`nodes.<node_id>.stats.system.board.gpus[].mem_total`	`integer`	Total GPU memory in GB.
`nodes.<node_id>.stats.system.driver`	`string`	Installed GPU or platform driver version.
`nodes.<node_id>.stats.system.machine_name`	`string`	Device host name.
`nodes.<node_id>.stats.cpu.mem_available`	`number`	Available system memory in GB.
`nodes.<node_id>.stats.cpu.mem_used`	`number`	Used system memory in GB.
`nodes.<node_id>.stats.cpu.mem_percent_used`	`number`	Percentage of system memory currently used.
`nodes.<node_id>.stats.cpu.cpu_percent_used`	`number`	Current CPU utilization percentage.
`nodes.<node_id>.stats.gpu[].type`	`string`	GPU model name.
`nodes.<node_id>.stats.gpu[].mem_total`	`integer`	Total GPU memory in GB.
`nodes.<node_id>.stats.gpu[].mem_available`	`number`	Available GPU memory in GB.
`nodes.<node_id>.stats.gpu[].mem_percent_used`	`number`	Percentage of GPU memory currently used.
`nodes.<node_id>.stats.gpu[].gpu_percent_used`	`number`	Current GPU utilization percentage.
`nodes.<node_id>.type`	`string`	Identifies the node type that produced this metadata.

JSON example

{
  "nodes": {
    "profile1": {
      "stats": {
        "cpu": {
          "cpu_percent_used": 0.0,
          "mem_available": 0.0,
          "mem_percent_used": 0.0,
          "mem_used": 0.0
        },
        "fps": 0,
        "gpu": [
          {
            "gpu_percent_used": 0.0,
            "mem_available": 0.0,
            "mem_percent_used": 0.0,
            "mem_total": 0,
            "type": "value"
          }
        ],
        "present": false,
        "process": {
          "running_pipelines_count": 0
        },
        "system": {
          "board": {
            "gpus": [
              {
                "mem_total": 0,
                "type": "value"
              }
            ]
          },
          "cpu_count": 0,
          "cpu_freq": 0.0,
          "driver": "value",
          "gpu_count": 0,
          "is_jetson": false,
          "lumeo_version": "value",
          "machine_name": "value",
          "mem_total": 0
        }
      },
      "type": "profile"
    }
  }
}