• EnglishEspañol日本語한국어Português
  • Log inStart now

NVIDIA Jetson integration

Our NVIDIA Jetson integration collects and sends data about your tegrastats utility to New Relic, where you can monitor CPU, GPU, RAM, and temperature readings about your NVIDIA Jetson device.

After setting up our NVIDIA Jetson integration, we give you a dashboard for your NVIDIA Jetson metrics.

Install the NVIDIA Jetson integration

Follow these steps to collect data about your NVIDIA Jetson device:

Install the infrastructure agent

To use the NVIDIA Jetson integration, you need to first install the infrastructure agent onto your host. The infrastructure agent monitors the host itself, while the integration itself extends your monitoring to include NVIDIA Jetson-specific data.

Save tegrastats to a file

From your shell environment, change to the tegrastats directory and run this command to save NVIDIA Jetson metrics at 5 second intervals:

bash
$
sudo tegrastats --interval 5000 --logfile tegrastats

Enable the NVIDIA Jetson integration with nri-flex

To set up the NVIDIA Jetson integration, follow these steps:

  1. Create a file named nvidia-jetson-config.yml and save it in the integrations directory:

    bash
    $
    sudo nano /etc/newrelic-infra/integrations.d/nvidia-jetson-config.yml
  2. Enable the agent to capture data from NVIDIA Jetson by adding the following snippet to your nvidia-jetson-config.yml file:

    integrations:
    - name: nri-flex
    config:
    name: JestonMetrics
    apis:
    - event_type: jetsonTegrastats
    commands:
    - run: tail -n 3 path/to/tegrastats
    split: horizontal
    set_header: [ram_used, ram_total]
    regex_match: true
    split_by: "RAM (\\d+)/(\\d+)"
    - run: tail -n 3 path/to/tegrastats
    split: horizontal
    set_header: [lfb_num, lfb_size]
    regex_match: true
    split_by: "\\(lfb (\\d+)x(\\d+)MB\\)"
    - run: tail -n 3 path/to/tegrastats
    split: horizontal
    set_header: [swap_used, swap_total, swap_cached]
    regex_match: true
    split_by: "SWAP (\\d+)/(\\d+) \\(cached (\\d+)MB\\)"
    - run: tail -n 3 path/to/tegrastats
    split: horizontal
    set_header: [iram_used, iram_total, lfb_block]
    regex_match: true
    split_by: "IRAM (\\d+)/(\\d+)kB\\(lfb (\\d+)kB\\)"
    - run: tail -n 3 path/to/tegrastats
    split: horizontal
    set_header: [cpu1_usage, cpu1_freq]
    regex_match: true
    split_by: "CPU \\[(\\d+)%@(\\d+)"
    - run: tail -n 3 path/to/tegrastats
    split: horizontal
    set_header: [cpu2_usage, cpu2_freq]
    regex_match: true
    split_by: ",(\\d+)%@(\\d+)"
    - run: tail -n 3 path/to/tegrastats
    split: horizontal
    set_header: [cpu3_usage, cpu3_freq]
    regex_match: true
    split_by: ",(\\d+)%@(\\d+)"
    - run: tail -n 3 path/to/tegrastats
    split: horizontal
    set_header: [cpu4_usage, cpu4_freq]
    regex_match: true
    split_by: ",(\\d+)%@(\\d+)\\]"
    - run: tail -n 3 path/to/tegrastats
    split: horizontal
    set_header: [emc_freq_pct, emc_freq_val]
    regex_match: true
    split_by: "EMC_FREQ (\\d+)%@(\\d+)"
    - run: tail -n 3 path/to/tegrastats
    split: horizontal
    set_header: [gr3d_freq_pct, gr3d_freq_val, ape, pll_temp]
    regex_match: true
    split_by: "GR3D_FREQ (\\d+)%@(\\d+) APE (\\d+) PLL@(\\d+)C"
    - run: tail -n 3 path/to/tegrastats
    split: horizontal
    set_header: [cpu_temp, pmic_temp, gpu_temp]
    regex_match: true
    split_by: "CPU@(\\d+\\.?\\d*)C? PMIC@(\\d+\\.?\\d*)C? GPU@(\\d+\\.?\\d*)C?"
    - run: tail -n 3 path/to/tegrastats
    split: horizontal
    set_header: [ao_temp, thermal_temp]
    regex_match: true
    split_by: "AO@(\\d+\\.?\\d*)C? thermal@(\\d+\\.?\\d*)C?"
    - run: tail -n 3 path/to/tegrastats
    split: horizontal
    set_header: [pom_in_currect, pom_in_peak]
    regex_match: true
    split_by: "POM_5V_IN (\\d+)/(\\d+)"
    - run: tail -n 3 path/to/tegrastats
    split: horizontal
    set_header: [pom_gpu_used, pom_gpu_total]
    regex_match: true
    split_by: "POM_5V_GPU (\\d+)/(\\d+)"
    - run: tail -n 3 path/to/tegrastats
    split: horizontal
    set_header: [pom_cpu_used, pom_cpu_total]
    regex_match: true
    split_by: "POM_5V_CPU (\\d+)/(\\d+)"

Find your data

You can use our pre-built dashboard template to monitor your NVIDIA Jetson data. Follow these steps to use our pre-built dashboard template:

  1. From one.newrelic.com, go to the + Integrations & Agents page, then click Dashboards
  2. Search for NVIDIA Jetson in the search bar, then select NVIDIA Jetson to install the dashboard.
  3. View your dashboard by going to one.newrelic.com > All Capabilities > Dashboards.

For docs on using and editing dashboards, see our dashboard docs.

Query your data with NRQL

With NVIDIA Jetson reporting data, you can also build queries to learn more about your device's performance:

  1. From any page in one.newrelic.com, click Query your data at the bottom of your browser.

  2. Build your query. For example, this is an example NRQL query to view the GPU temperature:

    SELECT latest(gpu_temp) as 'GPU temperature (C)' from jetsonTegrastats

What's next?

To learn more about building NRQL queries and generating dashboards, check out these docs:

Copyright © 2024 New Relic Inc.

This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.