Add support for per-process nvidia gpu memory usage.

[ash-shipit]

Description of the issue

• Currently, users cannot see gpu memory usage at a per-process level. They can only see gpu memory usage across the entire gpu. nvidia-smi does emit per-process gpu memory usage, but the cloudwatch agent does not parse that information.

Description of changes

• Extend current nvidia smi metrics to include per process gpu memory usage.

License

By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your choice.

Tests

• Added unit tests.
• Tested in personal account (build tested: linux/amd64/amazon-cloudwatch-agent.deb). See screenshots below. CloudWatch agent config used: https://pastebin.com/66jUu09T

Requirements

Before commit the code, please do the following steps.

1. Run make fmt and make fmt-sh
2. Run make lint

[github-actions]

This PR was marked stale due to lack of activity.