Moin,

thanks for the input! Some replies in-line.

On Fri, 2024-08-02 at 12:30 -0500, Mark Alley via mailop wrote:
> 
> Many of their customers use External Warning Tags, which inserts
> external warning text into the message body, obviously breaking the
> original DKIM body hash. (i.e. DKIM breaking is entirely dependent on
> the customer configuration) 

Ah, no, judging from the NDR, in this case, it is the urldefense
feature breaking DKIM.

> 
> Is the mail following this flow?  
>  
> [...]

Yep.

> If so, the Proofpoint-protected Google Workspace tenant owner needs
> to configure it (workspace) to not enforce email authentication
> policy, as SPF/DKIM auth will almost certainly be broken by the
> filter, and DMARC (where applicable) will fail for nearly all mail.

I sadly fear that the tenant falls into the not so small category of
"academic institution that clouded out all services and did not retain
sufficient in-house knowledge to be able to even have an email-addr. to
write to where such a request from an external party, i.e., me, would
be understood"; As such, getting them to fix things, is not really an
option.

> Odds are they're probably not receiving a lot of mail if this is the
> case. Another prime example of not following configuration
> requirements.

Based on my experience, though, I am rather sure that users in such
institutions are frequently wondering why so little mail actually makes
it through.

Tbh, when offering a service meddling with DKIM and SPF, I would expect
proofpoint to have some end-to-end monitoring in place to catch these
cases and notify corresponding customers. Setting something like that
up should not be the largest user story ever.

> SPF ~all would be the best case for maximum delivery with a strict
> DMARC policy. DMARC p=none would technically fix the problem, but
> obviously introduces more security concerns.

Added ~all for now; Let's see if that helps. As written in my follow-
up, though, it was already p=quarantine, so the message should not have
bounced/p=none should not make a difference for bouncing.

Anyway, let's see what ~all does to this issue.

> I'm sure you know this - but... pls no. 

I do; And I am not really serious... 

Then again, I need to get a (somewhat-ish) important email delivered,
and the sensible option you outlined (make the tenant fix their
settings) is not really feasible;

> Proofpoint doesn't seal ARC yet, unfortunately. Believe me, I've
> vehemently begged for it from their Product teams, as have many
> others.

Great, where do I get in line to join in on the begging... ? ;-)


With best regards,
Tobias
_______________________________________________
mailop mailing list
[email protected]
https://list.mailop.org/listinfo/mailop

Reply via email to