PG works okay with Peta Bytes of data. Various types of indexes, Table partitioning and parallel query computation can help.

Citus and other community extensions can help if you want to go distributed and fault tolerant