Connection Pool Stacking and query_wait_timeout Error with pgx v5.4.1 and PgBouncer

Nobody has claimed this yet.

I am facing an issue with the connection pooling behavior of the github.com/jackc/pgx package v5.4.1 in my Golang application. The application uses PgBouncer as a connection pooler between the Golang app and a PostgreSQL database. However, after deploying the application to my VPS and observing the logs, I noticed that the connection pool for my SELECT, INSERT, UPDATE, and DELETE queries keeps stacking up and eventually leads to a query_wait_timeout error when the connection limit is reached.

To troubleshoot this issue, I performed some investigations and found that downgrading the pgx package to v5.3.1 -- without even making any changes in the codes -- resolved the problem like magic✨. I deployed both the Golang app and PgBouncer using Docker containers.

**Expected Behavior**
The connection pooling should work correctly with pgx v5.4.1 and PgBouncer, efficiently managing and reusing connections without excessive stacking or timeouts.

**Actual Behavior**
The connection pool for queries keeps stacking up, leading to a query_wait_timeout error when the connection limit is reached.

**Version**
- Go: go 1.20
- PostgreSQL: PostgreSQL 13.11 (Debian 13.11-1.pgdg110+1) on x86_64-pc-linux-gnu, compiled by gcc (Debian 10.2.1-6) 10.2.1 20210110, 64-bit
- PgBouncer: image ```docker.io/bitnami/pgbouncer:1```
- pgx: v5.4.1

Start by reproducing the reported behavior with pgx v5.4.1, PostgreSQL 13.11, and the Dockerized PgBouncer setup, then compare it with pgx v5.3.1. Inspect connection-pool behavior around the query_wait_timeout and establish a minimal reproduction; done means identifying the version-specific cause and adding a verified fix or regression test.

Written by the indexing model from the issue text.