[Feature]: Option to Disable Metrics Health-Check Logging

[averyngo34]

🚀 The feature, motivation and pitch

Summary

Add a configuration option in vLLM to disable logging for metrics health-check endpoints.

Motivation

In production environments, frequent metrics health-check requests generate repetitive log lines that do not provide actionable value. This results in:

• Log clutter that makes it harder to identify real issues
• Increased log storage footprint and associated operational cost
• Higher noise-to-signal ratio when analyzing logs

Proposed Solution

Introduce a flag or configuration setting (e.g., --disable-metrics-healthcheck-logging or similar) that suppresses logs specifically for /health endpoints while keeping all other logging intact.

Expected Benefits

• Significant reduction in log volume in production
• Cleaner and more meaningful logs
• Lower storage and ingestion costs for centralized logging systems
• Improved user experience when debugging issues

Alternatives

No response

Additional context

No response

Before submitting a new issue...

• Make sure you already searched for relevant issues, and asked the chatbot living at the bottom right corner of the documentation page, which can answer lots of frequently asked questions.

[njhill]

Is disabling the uvicorn access logs altogether sufficient for this? (with --disable-uvicorn-access-log)

It will also suppress the per-request access logs but there is an option to enable vLLM native per-request logging that can be separately enabled.