Too many system events in QAN while monitoring Aurora MySQL

I’ve used PMM to monitor Percona XtraDB and MySQL instances before with little trouble. This is my first time using it to monitor Aurora MySQL. In QAN, I mostly see system events such as heartbeat and performance_schema queries. This happens until the performance_schema.events_statements_summary_by_digest table is full. According to AWS support, the size of this table cannot be increased on Aurora.

My question is, how do I “filter” the events logged in performance_schema to only show queries against user schemas?