Jim Seymour via Postfix-users <[email protected]> wrote:
> I don’t actually care what the TLS levels are, nor what the policies
> mean. I’m not interpreting them or acting on them—I’m simply
> reporting what is observed in the logs.
>
> So the question I’m trying to answer is this:
>
> What is a reasonable way to report what is seen in the logs, in a
> way that is useful and is faithful to what Postfix actually emits?
>
> What I’m currently inclined to do is something like:
>
> SMTP TLS Tokens Observed
> ------------------------
> nn level
> nn policy
> nn policy
> nn policy
> ...
> nn level
> nn policy
> nn policy
> nn policy
> ...
> ...
>
> That is: report how many times each level token is seen, and under
> each level, how many times each policy token is seen in conjunction
> with that level—without assigning semantics beyond what the log line
> itself provides.
>
> Does that seem like a reasonable and useful presentation of the data?
I am the maintainer of the FreeBSD port sysutils/logwatch and have patched the
postfix script in a way that it will just report whatever has been reported in
tls=mumble like:
11 Sent via SMTP ...
4 freebsd.org (tls=dane)
2 freebsd-current
1 freebsd-arm+confunsub-24770d5a4812114b
1 freebsd-arm+unsubscribe
3 gmx.de (tls=dane)
2 censored1
1 censored2
2 postfix.org (tls=dane)
2 postfix-users
1 sys4.de (tls=dane)
1 echo
1 univie.ac.at (tls=dane)
1 echo
Thus, any (tls=mumble) will become reported for any given target domain as is.
Next level are the individual mailboxes at those target domains.
I considered this level of detail as sufficient enough, and this patch has been
accepted upstream.
Regards,
Michael
_______________________________________________
Postfix-users mailing list -- [email protected]
To unsubscribe send an email to [email protected]
