I newly deployed the latest version as of this writing on Caprover which uses nginx as a reserve proxy for the containers.

Although MINIO_ROOT_USER and MINIO_ROOT_PASSWORD are set and correctly entered, and although MINIO_SERVER_URL is set correctly, MinIO panel says invalid login. Checking the browser console, it has HTTP 401 errors sometimes accompanied with HTTP 500.

This is caused by being behind a reverse proxy and the MinIO docs already have a configuration guide on the particular case of nginx. But if you are using this reverse proxy for serving other things too, like in the case of Caprover, or a similar self-hosted PaaS (like Coolify, Dokploy etc.), you wouldn't want to mess around the nginx config files. Instead, you could change the environment variable for the API URL to match the local domain instead of its public domain.

For example your public FQDN for the API is https://minio-api.example.com and this will be set to MINIO_SERVER_URL by default. Instead, use the local domain like http://minio-api.local-domain (along with the port if you need). (In my case of Caprover, it is something like http://srv-captain--minio-api) Mind that not only the domain changed but also the protocol did. I used http and not https, as it is locally served on the host.

Also on the API container, the UPSTREAM_HTTP_ADDRESS var should be local too. Instead of https://minio.example.com, in my case it is something like http://srv-captain--minio:9000. This resolves the invalid login error without having to edit any ngnix config.