fix(ci): rate limit and permission errors in trace publishing

[alisonshao]

Fix nightly trace publishing to correctly identify permission errors vs rate limit errors.

Previously all 403 errors were silently skipped as "rate limits", masking the actual issue: the GH_PAT_FOR_NIGHTLY_CI_DATA token lacks write permission to sglang-bot/sglang-ci-data.

Example failure: https://github.com/sgl-project/sglang/actions/runs/21051323020/job/60538104515#step:5:85

After this fix, permission errors will fail with a clear message instead of being silently ignored. The token secret will need to be updated with proper contents: write permission.

[gemini-code-assist]

Summary of Changes

Hello @alisonshao, I'm Gemini Code Assist¹! I'm currently reviewing this pull request and will post my feedback shortly. In the meantime, here's a summary to help you and other reviewers quickly get up to speed!

This pull request enhances the robustness of nightly CI trace publishing by accurately identifying and reporting GitHub API permission errors, which were previously masked as rate limit issues. This ensures that critical permission misconfigurations are no longer silently ignored, leading to clearer debugging. Additionally, a new tool has been integrated to provide insights into GitHub Actions runner utilization, aiding in resource management and optimization. A minor adjustment to a CUDA CI test suite is also included.

Highlights

• Enhanced Error Handling for GitHub API: Improved the publish_traces.py script to accurately distinguish between GitHub API rate limit errors and permission errors (both previously treated as 403s). Permission errors will now cause the CI job to fail loudly with a clear message, instead of being silently skipped.
• New Runner Utilization Report Script: Added a new Python script, runner_utilization_report.py, which analyzes GitHub Actions job data to calculate and report runner utilization metrics (active time, idle time, utilization percentage) per runner label. This report can be output in Markdown format.
• CI Test Suite Update: Modified test_eagle_infer_beta.py to update the CUDA CI test suite from 'stage-b-test-small-1-gpu' to 'stage-b-test-large-1-gpu'.

🧠 New Feature in Public Preview: You can now enable Memory to help Gemini Code Assist learn from your team's feedback. This makes future code reviews more consistent and personalized to your project's style. Click here to enable Memory in your admin console.

Ignored Files

• Ignored by pattern: .github/workflows/** (1)
  • .github/workflows/runner-utilization.yml

Using Gemini Code Assist

The full guide for Gemini Code Assist can be found on our documentation page, here are some quick tips.

Invoking Gemini

You can request assistance from Gemini at any point by creating a comment using either /gemini <command> or @gemini-code-assist <command>. Below is a summary of the supported commands on the current page.

Feature	Command	Description
Code Review	/gemini review	Performs a code review for the current pull request in its current state.
Pull Request Summary	/gemini summary	Provides a summary of the current pull request in its current state.
Comment	@gemini-code-assist	Responds in comments when explicitly tagged, both in pull request comments and review comments.
Help	/gemini help	Displays a list of available commands.

Customization

To customize Gemini Code Assist for GitHub experience, repository maintainers can create a configuration file and/or provide a custom code review style guide (such as PEP-8 for Python) by creating and adding files to a .gemini/ folder in the base of the repository. Detailed instructions can be found here.

Limitations & Feedback

Gemini Code Assist may make mistakes. Please leave feedback on any instances where its feedback is incorrect or counter productive. You can react with 👍 and 👎 on @gemini-code-assist comments. If you're interested in giving your feedback about your experience with Gemini Code Assist for Github and other Google products, sign up here.

You can also get AI-powered code generation, chat, as well as code reviews directly in the IDE at no cost with the Gemini Code Assist IDE Extension.

Footnotes

1. Review the Privacy Notices, Generative AI Prohibited Use Policy, Terms of Service, and learn how to configure Gemini Code Assist in GitHub here. Gemini can make mistakes, so double check it and use code with caution. ↩

[gemini-code-assist]

Code Review

This pull request effectively addresses the issue of distinguishing between rate limit and permission errors in the trace publishing script. The changes in publish_traces.py are clear and correctly implement the desired behavior.

However, this PR also includes two unrelated changes: the addition of a new runner_utilization_report.py script and a modification to a test suite name in test_eagle_infer_beta.py. For better repository history and more focused reviews, it would be ideal to submit these changes in separate pull requests. The new runner utilization script, in particular, is a significant feature that deserves its own PR with a detailed description.

Despite this, I have reviewed all the changes. My feedback includes a suggestion to refactor some duplicated code in publish_traces.py and several improvements for the new runner_utilization_report.py script related to type hinting and exception handling.

[gemini-code-assist]

This error handling logic for permission errors is duplicated at lines 453-459. To avoid code duplication and improve maintainability, consider extracting this block into a dedicated helper function. You could create a function like _handle_permission_error(repo_owner, repo_name) that prints the message and exits, then call it from both locations.

[gemini-code-assist]

Raising a generic Exception is generally discouraged as it can catch unexpected errors (e.g., KeyboardInterrupt). It's better to define and raise a more specific custom exception, like RuntimeError or a custom GhApiError, to allow for more precise error handling by the calling functions.

Suggested change

	raise Exception(f"gh api failed: {result.stderr}")
	raise RuntimeError(f"gh api failed: {result.stderr}")

[gemini-code-assist]

This function can return None, but the type hint only specifies datetime. To make it accurate, you should indicate that None is a possible return value. Using datetime | None is a good way to do this in modern Python.

Suggested change

	def parse_time(time_str: str) -> datetime:
	def parse_time(time_str: str) -> datetime | None:

[gemini-code-assist]

This function returns a list of dictionaries containing the utilization results, but it's missing a return type hint. Adding a type hint improves code clarity and helps static analysis tools.

Suggested change

	def calculate_utilization(repo: str, hours: int = 24, runner_filter: str = None):
	def calculate_utilization(repo: str, hours: int = 24, runner_filter: str = None) -> list[dict]: