docs: add MLflow AI Gateway documentation page

[PattaraS]

Summary

• Adds docs-website/docs/pipeline-components/generators/mlflowgatewaychatgenerator.mdx — a documentation page showing how to use OpenAIChatGenerator with MLflow AI Gateway
• Updates sidebars.js to include the new page in the generators section

What is MLflow AI Gateway?

MLflow AI Gateway is a database-backed LLM proxy built into the MLflow tracking server (MLflow ≥ 3.0). It exposes an OpenAI-compatible REST API (/gateway/openai/v1) that routes to OpenAI, Anthropic, Gemini, Mistral, Bedrock, Ollama, and more — configured through a UI with encrypted secrets management, fallback/retry, traffic splitting, and built-in usage tracing.

How It Works with Haystack

No extra packages needed — just use the existing OpenAIChatGenerator with api_base_url:

generator = OpenAIChatGenerator(
    model="my-endpoint-name",
    api_key=Secret.from_token("unused"),
    api_base_url="http://localhost:5000/gateway/openai/v1",
)

Page Contents

• Overview table (mandatory/output variables)
• Prerequisites: starting MLflow server and creating a gateway endpoint (UI, REST API)
• On its own usage example
• In a pipeline example
• Streaming example
• mlflow.deployments Python client alternative

Test Plan

• Page renders correctly in the docs site
• MLflow AI Gateway appears in the generators sidebar

This pull request was AI-assisted by Claude.

[vercel]

@PattaraS is attempting to deploy a commit to the deepset Team on Vercel.

A member of the Team first needs to authorize it.

[CLAassistant]

Thank you for your submission! We really appreciate it. Like many open source projects, we ask that you sign our Contributor License Agreement before we can accept your contribution.
_{You have signed the CLA already but the status is still pending? Let us recheck it.}