A team uses Amazon S3 Tables with Apache Iceberg and wants to reduce operational work for compacting files and managing old table snapshots and unreferenced files. Which capability should it evaluate?

1. Delete arbitrary Iceberg metadata files directly whenever the bucket contains many files.
2. Disable compaction permanently and assume Iceberg metadata alone merges every small data file during each query.
3. S3 Tables automatic table maintenance settings for supported compaction and retention cleanup.
4. Treat every noncurrent data file as safe to delete immediately, regardless of snapshot references.