[Bug]: LLM Proxy global advanced rate-limit policy does not apply globally unless keyExtraction is configured

[ashera96]

Please select the area the issue is related to

Area/Management (Management API or Management Portal UI)

Please select the aspect the issue is related to

Aspect/API (API backends, definitions, contracts, interfaces, OpenAPI)

Description

When advanced rate-limit policy is configured globally for an LLM Proxy, if keyExtraction is not set to apiname, a global bucket is not created; a resource-level bucket is initiated instead.

Note that this is working as expected for LLM Providers.

Steps to Reproduce

1. Create LLM provider and a proxy using it
2. Attach global rate-limit policy via the rate-limit configuring tab in the AI Workspace/ advanced-rate-limit policy via the Guardrails & Policies tab.
3. Deploy and invoke the LLM proxy and notice how the rate-limiting has not applied globally, rather unique buckets are created against each resource.
4. If keyExtraction is configured, this will work as expected by creating a global bucket

Severity Level of the Issue

Severity/Major (Important functionality is broken. Should be prioritized. Doesn't need immediate attention)

Environment Details (with versions)

No response

[ashera96]

This bug is currently exposed as default AI Workspace ratelimit is configuring an advanced-ratelimit policy. Once the fix for [1] is done, verify the default behavior when configuring global rate limit from AI Workspace UI works as expected and mark this issue as done. Proper fix for the underlying issue, which is if advanced-ratelimit policy is attached at the API-level, it should not be mandatory to define the keyExtraction to again specify that this needs to be an API-level policy; will be tackled via [2].

1. [Task]: Update the rate limit policy used for LLM Providers and Proxies from advanced rate limit to basic-rate limit #2338
2. [Bug]: Advanced Rate Limit not applying globally even when defined globally #2506