Question Spam detection on outgoing mails with Plesk Email Security

[Adarre]

Server operating system version

Ubuntu 22.04.5

Plesk version and microupdate number

Plesk Obsidian 18.0.66 #2

Hi,

How does spam detection on outgoing mails work on Plesk Email Security?

Spam is counted in the stats table. But how are they processed? For me, everything indicates that they are still delivered when a customer account is hacked. How can I make sure they are not delivered? There seems to be no possible setting...

Thanks

 

[Adarre]

I tested by sending myself spam-type emails to a second address. When I send myself the GTUBE character string, the email is blocked, I don't even receive it. But for everything else, the emails go through, with fairly low X-Spam-Scores. And yet my tests talk about viagra, money, with lots of red characters and emoji!!

Maybe I should directly modify the spamassassin config, but I'm afraid it will be overwritten by Plesk. In addition, apart from the required_score value, the rest I don't have the skills...

 

[Kaspar]

As you've already discovered mail is only blocked if the spam score hits the score threshold. You can either lower the mailbox threshold or customize (change, add, remove) any SpamAssassin rules or scores by adding them in separate config files in the /etc/spamassassin directory. Separate config files won't be overwritten by Plesk.

The default SpamAssassin configuration does a pretty poor job at recognizing spam. Adding your own custom spam rules and installing additional plugins (Pyzor, Razor, DCC, Spamhaus DQS to name a few) will improve the ability of SpamAssassin to recognize spam a lot.