Andy Pfister

judging from the tests, we simulate a `slow_close` and a `timeout` using Toxiproxy. more details are documented in #481 and in #445. Toxiproxy would also support to simulate a `TCPRESET`...

I also recently upgraded from a Debian-based Postgres v15 to pgautoupgrade v16, which is Alpine-based. As @rick-giorgi mentioned, I got some collation warnings (which did not go away), but I...

I would suggest `-bookworm` to align with Postgres (they also offer a `bullseye`-based image, which I think we do not need).

> We _**might**_ even be able to use the same pgautoupgrade script piece in both (via include file maybe?) which would be pretty optimal. 😄 maybe I do not understand...

So I took the most recent version of the entrypoint script from [here for Alpine](https://github.com/docker-library/postgres/blob/d08757ccb56ee047efd76c41dbc148e2e2c4f68f/16/alpine3.19/docker-entrypoint.sh) and [here for bookworm](https://github.com/docker-library/postgres/blob/d08757ccb56ee047efd76c41dbc148e2e2c4f68f/16/bookworm/docker-entrypoint.sh) and ran them through a text compare. There is one change...

> For our purposes we'll probably need to add some sample database (that uses extensions) to our git repo and use that for testing. I also had a similar thought...

sounds good, looking forward to it :)

@justinclift I modified the test script locally to load in a "pg-ified" version of the `AdventureWorks` example database by Microsoft (credits to https://github.com/lorint/AdventureWorks-for-Postgres). I initially wanted to write an example...

re: automatically upgrading extensions. I've looked into this a little bit as well during #71 - what I've seen is that `pg_upgrade` generates a SQL script with all the extensions...

@darkdecoy @justinclift I think we can close this issue, right?