[Feature] Adding raw metrics name conversion in metrics

[happyandslow]

Pull Request Description

Adding multi engine support by adapting different metrics exposed.

Related Issues

Resolves: #843

[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 introduces foundational changes to support metric collection from multiple inference engines. It refactors the metric retrieval logic to be engine-aware, allowing the system to correctly identify and fetch metrics even when different engines expose them under varying names. This enhancement is crucial for expanding compatibility and monitoring capabilities across diverse model serving environments.

Highlights

• Multi-Engine Metric Support: Implemented a mechanism to dynamically resolve raw metric names based on the inference engine type (e.g., vLLM, SGLang) associated with a pod. This allows the system to correctly identify and fetch metrics even when different engines expose them under varying names.
• Metric Definition Enhancement: The Metric struct in pkg/metrics/types.go now includes a RawMetricNameMapping field. This new map allows for defining different raw metric names for various engine types for a single logical metric, enabling flexible metric collection across diverse environments.
• Metric Fetching Abstraction: Introduced a new fetchMetrics helper function in pkg/cache/cache_metrics.go. This function centralizes the logic for retrieving engine-specific raw metric names by inspecting pod labels and utilizing the new metric mappings, replacing previous hardcoded metric name prefixes.

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 is currently in preview and 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 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

This pull request introduces multi-engine support for metrics by abstracting raw metric names through a mapping. This is a valuable enhancement for the project's extensibility. I've identified a critical issue regarding the handling of QueryLabel metrics that needs to be addressed.

[Jeffwan]

/lgtm

[Jeffwan]

please wait me to merge this one. it has > 20 commits and directly merge will leads to messy commit logs