Tiny enable soft watchdog in CI for stuck without logs

[fzyzcjy]

Motivation

cc @hnyls2002

Modifications

Accuracy Tests

Benchmarking and Profiling

Checklist

• Format your code according to the Format code with pre-commit.
• Add unit tests according to the Run and add unit tests.
• Update documentation according to Write documentations.
• Provide accuracy and speed benchmark results according to Test the accuracy and Benchmark the speed.
• Follow the SGLang code style guidance.
• Work with maintainers to merge your PR. See the PR Merge Process

[gemini-code-assist]

Summary of Changes

Hello @fzyzcjy, 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 addresses the challenge of identifying and diagnosing unresponsive processes, particularly in CI environments where logs might be absent. It introduces a robust and centralized watchdog mechanism that monitors the activity of critical system components. By integrating 'soft' watchdogs into the data parallel controller, detokenizer, and tokenizer managers, the system can now detect and report when these components become stuck, significantly improving debugging capabilities and system reliability without immediately terminating the process.

Highlights

• New Watchdog Utility: Introduced a new generic Watchdog utility (sglang/srt/utils/watchdog.py) to detect and handle processes that become unresponsive or 'stuck' without producing logs. This utility supports both 'soft' watchdogs (which log errors) and 'hard' watchdogs (which can terminate the process).
• Soft Watchdog Integration: Integrated the new soft watchdog mechanism into key manager components: DataParallelController, DetokenizerManager, and TokenizerManager. These watchdogs are configured to monitor the activity of their respective event loops and report if they become stuck.
• Scheduler Watchdog Refactoring: The existing watchdog logic for the Scheduler has been refactored to utilize the new generic WatchdogRaw class, centralizing the watchdog implementation and improving consistency across the codebase.
• Test Environment Variables: Added new environment variables (SGLANG_TEST_STUCK_DETOKENIZER, SGLANG_TEST_STUCK_DP_CONTROLLER, SGLANG_TEST_STUCK_TOKENIZER) to allow for deliberate testing of stuck states in CI, enabling verification of the watchdog's functionality.
• Watchdog Testing: New unit tests (test_soft_watchdog.py) have been added to the test suite to specifically validate the behavior of the soft watchdogs, ensuring they correctly identify and report stuck conditions in the detokenizer and tokenizer managers.

🧠 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.

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. ↩

[fzyzcjy]

/tag-and-rerun-ci

[gemini-code-assist]

Code Review

This pull request introduces a soft watchdog mechanism to detect stuck processes, which is a valuable addition for improving system robustness, particularly in CI environments. The implementation is clean, featuring a reusable Watchdog utility that has been integrated into the scheduler, data parallel controller, detokenizer, and tokenizer managers. The refactoring of the existing scheduler watchdog is well-executed. The new tests are also well-designed, using environment variables to simulate stuck components and verify the watchdog's functionality. I have one suggestion regarding the placement of the new test file in the test suite configuration.