This is exactly what I do. Debian + Postfix + rspamd.
It basically runs itself, I login to the rspamd web gui every now and then and add a few hosts to the "trusted DMARC" whitelist so they get a better score through rspamd, but I don't really need to do that.
I've been running it for ~5 years this way and it hardly ever lets a spam through (Very rarely the spams that are just images get through) and I've rejected 2 mails that I'm fairly sure were legit (just bulk marketing ones though)
People always say "Use Mailcow!" and I say "But why?". Adding the additional complexity of docker with the random firewall rules it adds, individual containers that need to be fed and watered. It's a mailserver, all the processes are doing the same mail related function.
Don't get me wrong, there's a very real place for Mailcow etc, but for a small end-user mailserver I think it adds overhead that makes running a mailserver more complex, not less.
I've been running it for ~5 years this way and it hardly ever lets a spam through (Very rarely the spams that are just images get through) and I've rejected 2 mails that I'm fairly sure were legit (just bulk marketing ones though)
People always say "Use Mailcow!" and I say "But why?". Adding the additional complexity of docker with the random firewall rules it adds, individual containers that need to be fed and watered. It's a mailserver, all the processes are doing the same mail related function. Don't get me wrong, there's a very real place for Mailcow etc, but for a small end-user mailserver I think it adds overhead that makes running a mailserver more complex, not less.
Anyway, great in depth article.