Beter SPAM configuration

[shoulders]

Each client account should have the ability to configure their spam rules.

Per Account (not per email address)

• Automatically Delete New Spam (Auto-Delete): - when the the email has a high enough score it is deleted.
• Email Filters - Allow me to identify SPAM emails by using custom regex
• Whitelist and blacklist
• Custom score - such things as a SPF fail can instantly be deleted.

Just tagging emails with ***INFECTED*** or ***SPAM*** is no good because the email still gets delivered and if that email goes to several places you would have to configure rules on each of them. I learnt this one from practice.

For reference

[24unix]

shoulder, just a suggestion from my side:

It's good when people show engagement, tell their ideas and want to improve an already good project.

BUT: Instead of flooding the forum with ideas/suggestions/whatever, for which at least 50% are already solved, you should spend one, two days exploring the Panel and its options.

In KH spam related stuff is handled via the webmail interface.

[Jolinar]

Just tagging emails with ***INFECTED*** or ***SPAM*** is no good because the email still gets delivered

That's the right way!
Mail filtering is the task of the client and not of the server...

[shoulders]

I appreciate what you are saying and I thank the community and KeyHelp staff for the feedback and their time.

In KH spam related stuff is handled via the webmail interface.

No one uses webmail any more unless it is for outlook or gmail. I know a couple of web agencies and they only forward mail or configure office 365.

Also rules set in the webmail would be email specific. I purposed account wide rules and filters.

Mail filtering is the task of the client and not of the server...

I don't agree. I do not want spam email on my phone all day long making it Beep 'you have mail' every 5 minutes. I have my email setup as POP on my desktop PC (leave on server for 14 days) and then via IMAP on my phone. SPAM emails are deleted before they even hit the mailbox. This combined with account level filtering means I get very little SPAM, ever.

With an account level REGEX (or two) I managed to stop five of my colleagues getting SPAM and I could keep on top of it by adding extra REGEX as needed. Most SPAM that is received has common links and formats even if they have different content. I did even consider adding the with a custom script examining emails in a honeypot email account.

[Jolinar]

I don't agree.

You are entitled to your own opinion.

BTW:
In one month, I receive about 50 spam emails, and my mail server doesn't filter anything; it just tags suspicious emails. I can't understand why some people have such problems with spam.

[24unix]

No one uses webmail any more unless

I never use webmail, but I also don't care about spam, I receive less than 10 spams a day on 10+ accounts, so why bother?
I mark them manually as spam and thats it.

it is for outlook or gmail

If that were the only companies offering eMail I'd have never in my live send or received an eMail.

Google is good for Searching, and that was it. MS is good for … erm, nothing. They don't exist in my world, and I'm not forced by my work to use any of their products.

[Jolinar]

Google is good for Searching, and that was it. MS is good for … erm, nothing. They don't exist in my world, and I'm not forced by my work to use any of their products.

Full ACK!

[Tobi]

In webmail you can configure all those filters you want. Then the server will also sort out the messages.

If you don't like working with webmail, just call it Spam Filter Central and it will work for your purposes.

[shoulders]

but it is not a central point, it is very isolated to each email address (and its aliases)

Will these rules persist if you change the webmail App?

I checked the filters and they are very basic.

[technotravel]

There must be a way to have Rspamd evaluate the mails and write the results into the mail header - haven't found out yet, how exactly to achieve this.

With these infos in the mail-headers, one could use the Sieve filters (in the webmail) - even to discard the mail entirely (if you trust the Rspamd evaluations). Then nothing will beep on your smartphone ...

But yes - this must be configured for each and every mailbox - as already mentioned, the handling of spam lies with the user. And yes - the Sieve filters are persistent, whatever webmail client you opt for (Roundcube or the other).

[shoulders]

I am going to install the new Debian 12 version of KeyHelp as this has a new way of handling SPAM. This might have the functionality I want.

I was using the Ubuntu version of KeyHelp for my current evaluation.

My thoughts: SPAM should be dealt with by the server and the administrators, never the user. The user wants to use his email without getting SPAM and get on with his real job.

[tab-kh]

My thoughts: SPAM should be dealt with by the server and the administrators, never the user. The user wants to use his email without getting SPAM and get on with his real job.

Maybe you mean "this user", not "the user". The server may reject a mail because of sender IP or technical reasons (SPF, DKIM, DMARC, ...). But once it has accepted the mail, it has to be delivered. There are most definitely a lot of users, which absolutely have to read every single mail which their mail server accepts with normal status code. Some of them even refuse to have spam moved to the junk folder. Their reasoning: The only difference would be, that I would have to check two folders instead of one.

Of course, every user is different, so this may not apply to you.