[FEAT] Add per-model token and error Prometheus metrics (part of #699)

[ardecode]

Summary

This PR adds model-level Prometheus metrics to improve observability in the router.

Metrics added

• vllm:input_tokens_total with labels [server, model]
• vllm:output_tokens_total with labels [server, model]
• vllm:request_errors_total with labels [server, model, error_type]

Implementation details

• Token usage is extracted from non-streaming responses.
• Error metrics capture exception types for easier debugging.

Scope relative to #699

This PR implements part of the metrics proposed in #699.

Not included (follow-ups)

• Per-model request count metrics
• Per-model latency histograms/summaries
• Resource or load indicators

Refs #699

[gemini-code-assist]

Summary of Changes

Hello @ardecode, 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 improves the observability of the router by introducing model-level Prometheus metrics. It enables detailed tracking of token usage for non-streaming requests and comprehensive error reporting, categorized by model and error type. This enhancement provides crucial insights into model performance and operational issues, forming a vital part of a larger initiative to bolster monitoring capabilities.

Highlights

• New Metrics Introduced: Three new model-level Prometheus metrics have been added: "vllm:input_tokens_total", "vllm:output_tokens_total", and "vllm:request_errors_total".
• Token Usage Tracking: Input and output token counts are now tracked for non-streaming requests, with labels for the "server" and "model".
• Error Tracking: Request errors are now captured and tracked, including the "server", "model", and the specific "error_type" (exception class name).
• Model Name Extraction: The "model" name is extracted from the incoming request body to provide granular metrics per model.

🧠 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 introduces valuable per-model Prometheus metrics for token usage and errors, which will significantly improve observability. The implementation is well-structured. I've identified a minor bug in the token counting logic that could lead to undercounting when a token count is zero and have provided a suggestion to fix it. Overall, this is a great addition.

[ardecode]

@zerofishnoodles ready for review!

[ardecode]

Hi @ruizhang0101 ready for review!

[ruizhang0101]

LGTM :)))