[Feature] Supporting new policies for xLLM

[happyandslow]

Pull Request Description

This PR creates two new policies based on engine metrics

• kv cache utilization: "least-gpu-cache"
• engine utilization (batch ratio): "least-utilization"

[gemini-code-assist]

Summary of Changes

Hello @happyandslow, 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 gateway's routing capabilities for xLLM by introducing two new metric-driven policies. These policies aim to improve resource balancing and efficiency by intelligently distributing requests based on real-time GPU cache and engine utilization metrics across the available pods.

Highlights

• New Routing Policy: Least GPU Cache: This PR introduces a new routing algorithm, least-gpu-cache, designed to route incoming requests to the xLLM pod that currently has the lowest GPU cache utilization. This aims to optimize resource distribution and prevent individual pods from becoming overloaded with cache usage.
• New Routing Policy: Least Utilization: A second new routing algorithm, least-utilization, has been added. This policy directs requests to the xLLM pod with the lowest engine utilization (batch ratio), promoting a more balanced workload across the available instances.
• Metric-Driven Routing: Both new policies leverage existing engine metrics (metrics.GPUCacheUsagePerc and metrics.EngineUtilization) retrieved from the system's cache to make informed routing decisions, enhancing the gateway's ability to intelligently distribute traffic.
• Fallback Mechanism: Each new routing algorithm includes a fallback mechanism. If valid metrics cannot be obtained for any pod, the system defaults to selecting a random pod from the available list, ensuring requests are still routed even under suboptimal metric conditions.

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 in your pull request via creating an issue comment (i.e. comment on the pull request page) using either /gemini <command> or @gemini-code-assist <command>. Below is a summary of the supported commands.

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 issue 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 or fill out our survey to provide feedback.

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

The code introduces new routing policies based on engine metrics. The review focuses on improving load distribution by handling ties in pod selection more effectively and removing a redundant nil check.

[Jeffwan]

the change looks good to me