Improve discoverability of OpenAI url service setting

The `openai` inference service accepts an optional `url` in `service_settings` to point at any OpenAI-compatible endpoint. Two issues:

1. The [REST API docs](https://www.elastic.co/docs/api/doc/elasticsearch/operation/operation-inference-put-openai) describe `url` as something that "can be changed for testing purposes." Can it be used in production for local LLMs and compatible gateways? Iff yes, the description should say that.

2. [Using OpenAI compatible models](https://www.elastic.co/docs/solutions/search/using-openai-compatible-models) covers this use case in detail but nothing links to it. Cross-referencing from the API docs or the inference processor page would help people find it.

Context: [elasticsearch#155700](https://github.com/elastic/elasticsearch/pull/155700) tried to add this info as a tip with a tracked marketing link. The info is already there, just hard to find.

Start with the REST API documentation for the OpenAI inference service and the “Using OpenAI compatible models” page linked in the issue. Confirm whether the optional service_settings url supports production use for local LLMs and compatible gateways, then add accurate wording and a cross-reference from the API docs or inference processor page. Done means the use case is clearly described and the existing guidance is discoverable.

Written by the indexing model from the issue text.