DeepStream metrics

[b.kowshik]

Please provide complete information as applicable to your setup.

• Hardware Platform (Jetson / GPU)
T4
• DeepStream Version
5.0
• JetPack Version (valid for Jetson only)
NA
• TensorRT Version
7.0.0.11
• NVIDIA GPU Driver Version (valid for GPU only)
440.64.00

• Issue Type( questions, new requirements, bugs)

• Are there any metrics that we can get from the gstreamer as well as the Nvidia plugins which are part of a DeepStream pipeline?
• Is there a list of such metrics?
• How can we get those metrics out of the pipeline into graphite/prometheus/datadog?
• Also what about GPU metrics, GPU-RAM metrics etc? What is the best way to get these metrics?

[Fiona.Chen]

@b.kowshik

Are there any metrics that we can get from the gstreamer as well as the Nvidia plugins which are part of a DeepStream pipeline?

For general gstreamer, there is no special metrics. If you want some, trace may be used although it will impact performance .Tracing (gstreamer.freedesktop.org). For Nvidia plugins, the latency of deepstream plugins can be measured but it also needs to be added by code. DeepStream SDK FAQ - Intelligent Video Analytics / DeepStream SDK - NVIDIA Developer Forums

@b.kowshik

Is there a list of such metrics?

No.

@b.kowshik

How can we get those metrics out of the pipeline into graphite/prometheus/datadog?

No. No tool to do this

@b.kowshik

Also what about GPU metrics, GPU-RAM metrics etc? What is the best way to get these metrics?

You can run “nvidia-smi dmon” to monitor the GPU loading and RAM usage in runtime.