feat: enhance one_logger_utils to support rankpulse integration

[sanshang-nv]

What does this PR do ?

Add support to new feature of one-logger in one-logger-utils package, which helps debug cpu/gpu hang problems. Be disabled as defaults and use several env vars to control:

• RANKPULSE_ENABLE to enable or disable rankpulse (default value is 0)
• RANKPULSE_INTERVAL_SECONDS to set checking period (default value is 15)
• RANKPULSE_TWINDOW_SECONDS to set history checking time window (default value is 300)
• RANKPULSE_GPU_DEBUG_INFO to enable or disable GPU debug information dump when GPU hang (default value is 1)

feature MR of one-logger repo: link

related integration MR of standalone one-logger-utils repo: link

Pre-checks:

• I have run the autoformatter.sh on my MR

[copy-pr-bot]

This pull request requires additional validation before any workflows can run on NVIDIA's runners.

Pull request vetters can view their responsibilities here.

Contributors can view more details about this message here.

[pablo-garay]

I am not the best / with most context for this, please have the other reviewers tagged here approve it

[sanshang-nv]

I am not the best / with most context for this, please have the other reviewers tagged here approve it

thanks, @pablo-garay . Could you please help AT the right reviewer to help approve if you know? thanks!

[pablo-garay]

I am not the best / with most context for this, please have the other reviewers tagged here approve it

thanks, @pablo-garay . Could you please help AT the right reviewer to help approve if you know? thanks!

I thought the tagged reviewers right now should/would suffice? E.g. a few of them reviewed/commented already - hence Was in their review would be enough

[sanshang-nv]

kindly ping @jaredcasper @deepakn94

[jaredcasper]

Are we using environment variables? We don't use environment variables for anything else. Can you use command line arguments?