Improving filtering of SPAM...

Patrick Ben Koetter p at sys4.de
Thu Jun 20 23:22:22 CEST 2013


* Christopher Koeber <ckoeber at gmail.com>:
> Now, the current install isn't completely ineffective; it catches a great
> deal of spam messages but my question stems from asking what can I do to
> greatly increase the accuracy.
> 
> Is there a guide somewhere on improving detection of spam with AmavisD? I
> couldn't find such a guide. Much of the items mentioned for improving
> SpamAssassin was within the "local.cf" file but I know AmavisD ignores that
> (from what I read). I also know I can lower the tag/tag2/etc. numbers but
> then valid mail gets caught.

Just a little nitpicking: amavisd doesn't detect spam; spamassassin does.
amavisd only provides the content framework around various filters including
spamassassin.

If you catch lots of spam you may, as a very basic measure, insure your
filters have an understanding of what internal and external clients are and
what inbound and outbound traffic is. If they don't then you should fix that
because filters give different scores depending on direction and sender.
You may, if you can read German, read my blog entry on that:
<http://sys4.de/de/blog/2013/02/15/e-mail-content-policies-mit-amavis/>

Additionally, if you use whitelisting, either train and maintain it or turn it
off. There's nothing more annoying than an automatism that automatically does
the wrong thing.

For further, and better debugging, I suggest you take a look at the spam
report headers in messages scanned by SA. Try to find out WHY they fired and,
if they do the wrong thing, WHAT make them do the wrong thing.

Pretty basic stuff, but almost everything you are going to optimize builds
upon that.

p at rick


-- 
[*] sys4 AG
 
http://sys4.de, +49 (89) 30 90 46 64
Franziskanerstraße 15, 81669 München
 
Sitz der Gesellschaft: München, Amtsgericht München: HRB 199263
Vorstand: Patrick Ben Koetter, Axel von der Ohe, Marc Schiffbauer
Aufsichtsratsvorsitzender: Florian Kirstein
 


More information about the amavis-users mailing list