A producer sends JSON records to Amazon Data Firehose for delivery to S3. Analysts want Parquet files and can tolerate delivery buffering. Which managed feature most directly meets the requirement?

1. Change the S3 object key suffix to .parquet while leaving the object contents as JSON.
2. Create only a Glue Data Catalog table for the JSON objects and assume catalog registration rewrites them as Parquet.
3. Configure Firehose record-format conversion using an appropriate AWS Glue Data Catalog schema.
4. Enable S3 Versioning and expect it to convert every JSON object to Parquet.