A data engineer suspects that a query is suffering from poor performance because Spark is failing to utilize predicate push-down. Where in the Spark UI can they definitively diagnose whether predicates are being pushed down to the data source?