spam and the BTS
My crossassassin test seems reasonable so far. About 1/4 (14 out of
62 if I recall correctly) of the spam not caught by spamassassin had
scores 4 or more, and none of them had a spamassassin score <= 0.
Of the 60 or so non-spam messages that had a crossassin score >= 4,
all had spamassassin scores < 0.
I'll continue in test mode for now, it's not a dramatic improvement.
Looking into the spamscan log, I find that the problem was spamscan
processing messages slowly rather than unusual bursts of spam.
Durring some of the major backlogs, it was processing as few as 26
messages in 5 minutes. My guess is that there was a denial of service
attack against one or more of the DNSBLs being used.
(~blarson/bin/sslog produced ~blarson/sslog.log if anyone else wants
to look at the details.)
Blars Blarson email@example.com
With Microsoft, failure is not an option. It is a standard feature.