[Iceberg] Add remote benchmarks

Part of #599.

Add repeatable remote benchmarks for the Iceberg integration.

Scope:

• Extend the remote benchmark environment to register Iceberg tables from object storage.
• Document required object-store configuration and benchmark invocation.
• Run representative Iceberg workloads against a remotely accessible table.
• Establish a baseline for comparing local and remote execution.

Ideally, the same datasets that are supported for raw parquet files should also be supported for Iceberg.

They probably will have a different shape, and as the remote benchmarks are designed to run on AWS, they will probably need to be wired up to AWS Glue or something.

As this requires an expensive infra setup, this should be handled by a Datadog employee first, even if the remote setup should work for everyone at the end of the day.