Nginx strange upstream address after app restart

I have an application running in a container with an exposed port 8082. There is also an nginx configuration, part of which is shown below. In a normal situation, I see upstream: 127.0.0.1:8082 in the access log, which is correct. But when I stop the application container and start it again, I see to: localhost in the logs for about 5-6 seconds, which leads to a 502 error on the site. At this time, the application is already live and running on port 8082. Why does this happen?

server {
    server_name acme.com;

    access_log /var/log/nginx/acme.log upstreamlog;

    location / {
       proxy_pass http://localhost:8082;
       proxy_next_upstream error timeout http_502;
       proxy_set_header Host $host;
       proxy_set_header X-Real-IP $remote_addr;
       proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
       proxy_set_header Upgrade websocket;
       proxy_set_header Connection Upgrade;
    }


    listen 443 ssl; # managed by Certbot
    ... some other Certbot lines ...
}

> sudo nginx -T | grep acme.com
nginx: the configuration file /etc/nginx/nginx.conf syntax is ok
    server_name acme.com;

Nowadays, localhost means both the IPv4 127.0.0.1 and the IPv6 ::1. When NGINX forwards traffic to the upstream, it takes the host name you gave which yields both IPv4 and IPv6, while I'm guessing your upstream is IPv4 only.

It seems I managed to solve the problem by replacing proxy_pass http://localhost:8082; with proxy_pass http://127.0.0.1:8082;.