[New Feature] APIM deployment scenario for general LLMs deployed in Azure AI Studio

[sushaanttb]

What?
Proposal to add a new deployment scenario for LLMs deployed in Azure AI Studio

Why?
Until now APIM supported Azure OpenAI as the only GenAI backend along with Azure OpenAI specific [policies].(https://learn.microsoft.com/en-us/azure/api-management/azure-openai-emit-token-metric-policy)

However, as per the following blog posts:

1. https://techcommunity.microsoft.com/blog/aiplatformblog/expanding-the-azure-ai-model-inference-api-integrating-azure-ai-into-your-develo/4212883
2. https://techcommunity.microsoft.com/blog/integrationsonazureblog/expanding-genai-gateway-capabilities-in-azure-api-management/4214245

We have following new policies now available in APIM:

1. LLM Token Limit Policy
2. LLM Emit Token Metric Policy
3. LLM Semantic Caching Policy

essentially making APIM to now support any LLM deployed via Azure AI Studio (yay!)

Hence, a strong reason in itself for us to now also add deployment scenario on same i.e. using any LLM model from Azure AI Studio model catalog as the APIM Backend with these new policies.