Publishing GPU metrics to Oracle Cloud Infrastructure (OCI) Monitoring service
Currently, the Oracle Cloud Infrastructure (OCI) Monitoring service does not have built-in support for collecting GPU metrics from GPU instances.
However, it's possible to publish custom metrics to OCI Monitoring service. This repo has the necessary information and the shell script for publishing GPU temperature, GPU utilization, and GPU memory utilization from GPU instances to OCI Monitoring service.
If you encounter any problems when following this guide, feel free to create an issue.