Target file size is dynamic based on the size of the table.  See:
https://docs.databricks.com/en/delta/tune-file-size.html#autotune-file-size-based-on-table-size

Row Level Concurrency helps avoid conflicts when two operations are working on rows within in the same file, but not the same row.  In a concurrent append exception, new data is added while a merge/update/delete is also running on the table, RLC doesn't help this.