Presently, if the Docker daemon is required and isn't found, deploys continue rather than aborting, which often leads to orphaned resources that we need to remove individually (usually either a server function or the DNS records for a server function). If we abort deploys with Ctrl+C as soon as we see that error, though, we're usually fine.

It would be great if we could opt into SST doing this itself when it encounters this error, as it's really common to forget to start Docker before the first deploy of the day, especially if you don't have it start on login.