We use latest for everything from this repo, but I have noticed this behaviour in the past, so it's not a regression.

I have witnessed the problem on two different machines:

Docker version 20.10.7, build f0df350
Docker version 20.10.6, build 370c289

We use the three container Docker Compose solution described in the documentation.

Bug description

When the server (or the Docker daemon) is restarted, then while all containers come up correctly, many, sometimes all 'external' containers are not proxied to correctly. To be clear, I am talking about third party containers that 'request' proxying through this repos services by having the appropriate environment variables set. Restarting this repos containers, and then restarting the problematic containers afterwards, results in correct proxying, so I suspect that the problem is that the order in which containers are starting back up.

acme-companion image version

Info: running acme-companion version v2.1.0-16-g027c3cf

nginx-proxy configuration

We cloned this repo at 027c3cf3

The other containers are a variety of different services, none of which configurations would matter, I would expect, as they all have different configs and all suffer from the same problem. As there are about ~80 different ones, unless there is anything specific, I'd rather not spend a few hours sanitizing and posting the different configs ;)

Docker host

• OS: CentOS 8, Raspbian GNU/Linux 10 (buster)
• Docker version 20.10.7, build f0df350, Docker version 20.10.6, build 370c289