Read replica lag becomes proportionally high when writing to Aurora postgres master instance

Question

We currently use a Scala spark app that fires up copy commands to the Aurora Postgres write instance.

The pipeline writes around 5Tb of data every day with different part