Hi again,
innodb_log_file_size should be tuned to hold around 1 hour worth of write traffic at peak load as explained here: How to calculate a good InnoDB log file size - Percona Database Performance Blog
depending on how big your ETL/data load size is, you might require a bigger than 1 GB redo log
If you are on a cloud provider, it might be possible that after running at peak load you run out of burst credits: Understand Amazon EFS burst credits
Double check that write load is not spiking/saturating all of a sudden and for a long period.
if dataset/operations increase monotonically over time, after certain threshold you reach some limit/boundaries and performance degrades.
it is strongly suggested to setup a monitoring system in place (such as PMM free and open source which you can take a look here Grafana ) so that you can check how MySQL behaves and where the bottlened/satured subsystem is