Skip to content

[Docs] Add documentation for show_metrics(), stream_logs(), and job notification setup#6065

Merged
mujtaba1747 merged 3 commits into
aws:master-nova-follow-upsfrom
ehsu3:docs/job-monitoring-notifications
Jul 21, 2026
Merged

[Docs] Add documentation for show_metrics(), stream_logs(), and job notification setup#6065
mujtaba1747 merged 3 commits into
aws:master-nova-follow-upsfrom
ehsu3:docs/job-monitoring-notifications

Conversation

@ehsu3

@ehsu3 ehsu3 commented Jul 20, 2026

Copy link
Copy Markdown
Contributor

Issue #, if available: N/A

Summary

Adds documentation for show_metrics(), stream_logs(), and job notifications to the existing Sphinx RST pages under model customization.

Changes

File What was added
docs/model_customization/model_customization.rst "Job Notifications" under Key Features — SNS notification setup, config options, rule management, prerequisites, and deduplication behavior
docs/model_customization/finetuning_serverful.rst Updated the "Monitoring" section — show_metrics() and stream_logs() examples for SMTJ compute
docs/model_customization/finetuning_hyperpod.rst Updated the "Monitoring" section — show_metrics() and stream_logs() examples for HyperPod compute, including Ctrl+C behavior note

Testing

  • make html in docs/ builds without errors
  • Verified rendered pages

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


# After a kernel restart, use the standalone function with job name
from sagemaker.train import plot_training_metrics
plot_training_metrics("my-sft-job-20260101195119")

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

nit: maybe a more basic job name e.g. my-sft-job

Copy link
Copy Markdown
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I can update that!

df = trainer.show_metrics(metrics=["training_loss", "lr"])

# Filter by step range
df = trainer.show_metrics(starting_step=20, ending_step=100)

Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

is show_metrics supported in serverless?

job = trainer.train(wait=False)

# Stream logs directly to the terminal (blocking)
trainer.stream_logs()

Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

And is stream_logs() supported in serverless?

Copy link
Copy Markdown
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

They are both supported for serverless, but we don't have an equivalent finetuning_serverless.rst. Syncing offline to figure out the best place to put this information.

@mujtaba1747
mujtaba1747 merged commit fbbc69a into aws:master-nova-follow-ups Jul 21, 2026
1 check passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants