As a platform admin, I want to define and enforce rate limits on tool usage per user/tenant, so that I can ensure fair resource distribution across consumers or allow for tiered usage based on plans.

Considerations:

• Define limits per User + Tool combination
• Track number of calls & provide usage metrics (e.g. Prometheus or OpenTelemetry) for observability
• Is there a standard response in MCP, or should response be standard HTTP 429?
  • Allow specifying the response content & code?
• Counters should be time-window based (e.g. per minute/hour/day)

The existing Kuadrant RateLimitPolicy can be leveraged, similar to how it was demoed in the proof of concept https://github.com/Kuadrant/Kuadrant-MCP-Walkthrough