feat: Enable Ray dashboard for Ray state API

[pjin-nvidia]

What does this PR do ?

This PR sets a non-localhost dashboard address when starting the Ray cluster via ray.sub to enable queries to the Ray state API (potentially used by e.g. Gym).

Issues

List issues that this PR closes (syntax):

Usage

• You can potentially add a usage example below

# Add a code snippet demonstrating how to use this

Before your PR is "Ready for review"

Pre checks:

• Make sure you read and followed Contributor guidelines
• Did you write any new necessary tests?
• Did you run the unit tests and functional tests locally? Visit our Testing Guide for how to run tests
• Did you add or update any necessary documentation? Visit our Document Development Guide for how to write, build and test the docs.

Additional Information

• ...

Summary by CodeRabbit

• Chores
  • Ray head node dashboard now binds to the head node IP address and is explicitly enabled in the start command.

_{✏️ Tip: You can customize this high-level summary in your review settings.}

[coderabbitai]

📝 Walkthrough

Walkthrough

Two flags are added to the head node Ray start command: --dashboard-host="$head_node_ip" and --include-dashboard=True. These explicitly configure dashboard exposure by setting the host IP and enabling dashboard inclusion.

Changes

Cohort / File(s)	Change Summary
Ray head node dashboard configuration ray.sub	Add --dashboard-host="$head_node_ip" and --include-dashboard=True flags to head node Ray start command

Estimated code review effort

🎯 1 (Trivial) | ⏱️ ~3 minutes

• Focus areas: Verify the flag syntax is correct for the Ray version in use and confirm the head node IP variable ($head_node_ip) is properly defined and populated in the script context.

Pre-merge checks and finishing touches

✅ Passed checks (4 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title accurately reflects the main change: enabling the Ray dashboard by adding dashboard-related flags to the head node Ray start command for Ray state API access.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.
Test Results For Major Changes	✅ Passed	PR contains only minor configuration changes (two flag additions) that qualify for exemption from test documentation requirements per custom check instructions.

✨ Finishing touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Post copyable unit tests in a comment
• Commit unit tests in branch pjin/ray-state-api

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[coderabbitai]

Actionable comments posted: 1

📜 Review details

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro

📥 Commits

Reviewing files that changed from the base of the PR and between a99bc26 and bea6d2b.

📒 Files selected for processing (1)

• ray.sub (1 hunks)

🧰 Additional context used

📓 Path-based instructions (1)

!(**/tests/**|**/test_*.py|**/test_*.sh)

📄 CodeRabbit inference engine (CODING_GUIDELINES.md)

Add the NVIDIA copyright header to all Python files and shell scripts (excluding tests). The header should include the current year

Files:

• ray.sub

⏰ Context from checks skipped due to timeout of 90000ms. You can increase the timeout in your CodeRabbit configuration to a maximum of 15 minutes (900000ms). (3)

• GitHub Check: Lint check
• GitHub Check: Post submodule check comment / Comment on PR
• GitHub Check: Post automodel integration comment / Comment on PR

[terrykong]

@pjin-nvidia is this so you can access using DNS of the slurm node and not have to hop onto and ssh port forward?

[pjin-nvidia]

@pjin-nvidia is this so you can access using DNS of the slurm node and not have to hop onto and ssh port forward?

yeah, also it makes it easier to use ray.util.state APIs (e.g. https://docs.ray.io/en/latest/ray-observability/reference/doc/ray.util.state.list_nodes.html#ray.util.state.list_nodes) which I think work by querying the dashboard