Docker deployer should attempt to reattach to lost containers #6281
Description
From: #6265 (comment)
it might be confusing UX wise that if a container exits (in my test via os.Exit(1) & docker kill ) skaffold does not give any information to the user, the logs just stop. I think a way to do this would be to open a new ContainerWait and wath for a status change after containers are running:
https://pkg.go.dev/github.com/docker/docker/client?utm_source=godoc#Client.ContainerWait
We should implement a way for Skaffold to attempt to reattach to a lost container, and additionally emit an event and a warning to users that a container has been lost. This should also remove the container from being tracked so Skaffold doesn't try (and fail) to clean it up later.