ES|QL Data Federation: Query Apache Iceberg tables

## Value Proposition

Many enterprises organize their analytical data as Iceberg tables layered over object storage. Unlike raw file queries in [Data Federation](https://github.com/elastic/roadmap/issues/307) where schema must be inferred, Iceberg provides rich metadata — schema, partition layout, snapshot history — that enables more efficient query planning and eliminates schema inference guesswork.

## Expected Outcome

ES|QL Data Federation supports querying Apache Iceberg tables as datasets. The query planner uses Iceberg metadata for schema resolution and partition pruning, enabling zero-copy analytics on managed data lake tables without ingestion.