I am using Postgres engine 16.3 on RDS and can successfully connect to and run my program.

My program is a batch process as part of ETL and processes sequential data from JSONL file.

In increased the number of threads doing the same work (SELECT and some INSERT) and sometimes get the following error:

no pg_hba.conf entry for host

It seems as though the thread count breaches the CPU limit, and after a short allowable burst, it starts declining queries. Lowering the thread count has has had a positive effect.

But sometimes, after an hour or so I might get the odd error. Which is super annoying.

Given how my process is built, it means I need to restart that particular file.

What causes this intermittent error and how can I confidently prevent it if I am going to run a batch for upwards of 24 hours (first run).