I'm completely confused about question of necessity/benefit of stripping local network host names and IPs from from mail headers.
Two years ago I've found an article which states that modern ESPs (Yahoo, Microsoft, Google...) considers Received: from localhost [192.168.0.100] ... like headers somehow related to spamyness of email, and gives email bad treatment consequently. From then, all my Postfix servers strips that info with header_checks directive.
Recently I've come across KBs that states opposite - that, in short, every header MTA generates is useful and should not be removed forcibly.
For now, I'm not experiencing issues, but these two schools cannot be right at the same time - do you know what practice is the best regarding this issue?
This will boil down to opinion in one of two camps - one in favor of hiding that information, the other in favor of having it. Both camps will present valid arguments in their favor, but ultimately it is much more an opinion-based decision. I tend to favor hiding internal information when possible, with the justification that removing the internal headers helps hide the (proprietary) internal network structure from potential intruders. I don't believe there is a definitive "best practice" on this issue.