On Mon 22/Jul/2024 20:05:31 +0200 Brotman, Alex wrote:
We send reports from a platform that doesn't contribute to DMARC reports, as 
option 4.


Me too, I send them as [email protected]. This is a different problem, though. I think both problems should be addressed in the spec.


1) What if they fix it?


Could make them expire after a month or so?


3) I don’t think this is a valid option.  You could have millions of messages 
from a single IP.  Perhaps you mean omit reports for a single message (or below 
some other threshold)?


In any case, (3) tends to give a partial view. I mean, I know that reports about less than hundreds are just noise for ESPs. However, in theory, mail can also be used by the average housewife...


However, if it's a bounce (I take that to mean a permanent failure), it won't 
always result in a DMARC report.


Andreas' NOTIFY=NEVER seems to be the correct solution to this problem.


Best
Ale
--

-----Original Message-----
From: Alessandro Vesely <[email protected]>
Sent: Monday, July 22, 2024 1:31 PM
To: dmarc-ietf <[email protected]>
Subject: [dmarc-ietf] Aggregate reporting loops

Hi,

non-existing rua= addresses generate loops, because the target domain sends
a bounce, and on the next day the generator sends them a report for that one
message.  The report bounces, and so forth...

Three ways to prevent that:

1, accurately store all bouncing addresses so as to avoid sending again,

2, omit aggregating data for DSNs, or

3, omit sending reports that have only one row.

_______________________________________________
dmarc mailing list -- [email protected]
To unsubscribe send an email to [email protected]

Reply via email to