[model-gateway] add retry support to OpenAI router chat endpoint

[slin1237]

• Add RetryConfig field and initialize from router config
• Wrap route_chat request execution with RetryExecutor
• Use is_retryable_status for 408, 429, 500, 502, 503, 504 status codes
• Record circuit breaker outcomes correctly:
  • Failure on connection error or error status
  • Success only after body read (non-streaming) or stream start (streaming)
• Record retry metrics with BACKEND_EXTERNAL label
• Record duration/error metrics after retry completes
• Responses endpoint not included (per requirement)

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 @slin1237, 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 significantly enhances the robustness of the OpenAI router's chat endpoint by introducing a comprehensive retry mechanism. It leverages a RetryExecutor to automatically reattempt requests that encounter transient network issues or specific server-side errors, ensuring greater reliability. The changes also refine the circuit breaker and metrics reporting to accurately reflect the outcomes of these retried operations, providing better observability into the system's resilience.

Highlights

• Retry Configuration: Added a RetryConfig field to the OpenAIRouter and initialized it from the router's configuration.
• Retry Execution: Wrapped the route_chat request execution with a RetryExecutor to handle transient failures automatically.
• Retryable Status Codes: Configured the system to retry requests for specific HTTP status codes: 408 (Request Timeout), 429 (Too Many Requests), 500 (Internal Server Error), 502 (Bad Gateway), 503 (Service Unavailable), and 504 (Gateway Timeout).
• Circuit Breaker Logic: Improved the circuit breaker logic to record failures on connection errors or non-success HTTP status codes, and successes only after the full response body is read (for non-streaming) or the stream has successfully started (for streaming).
• Metrics Integration: Integrated new metrics for retries, including the BACKEND_EXTERNAL label, and adjusted existing duration/error metrics to be recorded after the entire retry process completes.
• Scope Limitation: The route_responses endpoint was explicitly excluded from this retry functionality, focusing the changes solely on the chat endpoint.

🧠 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. ↩

[gemini-code-assist]

Code Review

This pull request effectively adds retry support to the OpenAI router's chat endpoint. The use of RetryExecutor encapsulates the retry logic cleanly. I appreciate the careful handling of circuit breaker state updates for both streaming and non-streaming responses, ensuring success is only recorded after the response is confirmed. The addition of retry-specific metrics is also a great enhancement for observability. I've left one comment regarding a performance optimization to avoid expensive cloning within the retry loop. Overall, this is a solid implementation of a critical reliability feature.