Interestingly, is this a pressure test? Can you please elaborate a bit more on how the application tests it?
It was a custom coded application in Rust designed to demonstrate how this feature can reduce connections in the connection pool (deadpool of Rust with 16 connections in the pool), Because primary can commit asynchronously and release the connection back to pool and then wait for standby. I don't mind sharing the code, if that helps.
A client side workaround similar to the following is a temporary solution for me for my tests:
...
I think we better fix the page/segment header issue on the server side.
Agree, +1
We cannot expect every application to impliment workaround after documenting the known problem
