Hello. SpamSieve has been working well for 20 days since installed and configured. Then suddenly, with 10 days remaining on the trial and a favorable opinion, it suddenly stopped working.
Details: I restarted macOS around 10 PM last night (just a routine I do about once a week that helps keep macOS running smoothly). Everything appeared to be fine when I went to bed around 11:30 PM. This morning, I woke up to find 30+ spam messages in my inbox. The first spam started appearing in my inbox at 12:09 AM this morning and is steadily incoming ever since.
What’s the best way to troubleshoot this? I can restart SpamSieve and Apple Mail to see if that works, but first—what should I do with all the spam? A lot if it is junk that was already being successfully filtered until 12:09, so I’m guessing I don’t need to retrain all 30+ messages as spam. Would it hurt? More importantly, what if restarting doesn’t fix the problem?
Update: I just remembered to check the log and found a lot of spam suddenly being auto-trained as good. See a few examples below. These are definitely junk messages that were being filtered as spam up until last night.
P.S.: Running macOS: 13.2.1 (installed last week without any immediately apparent issues)
A sudden change like that usually means that SpamSieve was simply turned off in Mail. So you should check the setup and (if it already looks correct) see whether those messages are mentioned as Predicted: Goodin the log.
If the messages were Predicted: Good, they are SpamSieve mistakes and should be trained as spam. If they were not predicted (because of a setup problem) then you can instead select them and choose Message ‣ Apply Rules to tell SpamSieve to refilter them.
The log excerpt above shows that these two messages were predicted to be good because the sender matched your whitelist. This likely means that previous spam messages had been Predicted: Good but that you did not train them as spam, so SpamSieve continued to think they were good.
Thanks again for the quick responses. This along with the 30-day trial gives me confidence to make the purchase.
Everything checks out okay as far as setup.
I’m still baffled by messages marked as “Trained: Good (Auto)”. For example, email “From: Lume” has been automatically sent to junk for weeks now. And I definitely didn’t train it (or anything else) as good last night. In addition to Predicted Good, if I manually retrain Trained Good items as spam, will SpamSieve clean up its whitelist or will it leave the old rule in-place and add a conflicting new rule?
It’s normal for them to be Trained: Good (Auto) if they were Predicted: Good. The message in the screenshot was Predicted: Good because of a previous (uncorrected) message from Lume Deodorant.com Ad Partner. You could look farther back in the log to see why that one was Predicted: Good. Or if you send in a diagnostic report I can investigate.
Well, there shouldn’t be any incorrect Trained: Good messages that were not also Predicted: Good unless you trained them by accident. Yes, it’s essential that you train all the Predicted: Good messages that are not actually good. This will disable the old rule on the whitelist and add a new rule on the blocklist. There’s more information about that under Spammy Whitelist Rules.
It turns out the sudden junk email spike was caused by Outlook. I created many spam rules in Outlook over the years and wanted to leave most of them active while letting SpamSieve catch whatever made it through the cracks.
After you gave me a better understanding of how SpamSieve functions and how to ensure it was running properly, I started to suspect an Outlook junk filter outage last night. Obviously, SpamSieve wouldn’t be able to detect all the new spam it had never been trained on.
In my experience after you faithfully train SpamSieve for a short period of time, it is virtually plug and play. Occasionally there will be a new bunch of Spam, but it is easy to get control on those uncommon situations.
“In my experience after you faithfully train SpamSieve for a short period of time, it is virtually plug and play.”
The term Plug and Play refers to hardware peripherals that you can simply plug in and start using because the system automatically recognizes it and installs the correct drivers, settings, etc. Sometimes the term is loosely used for apps as well. So, the initial setup and training steps for SpamSieve makes it not plug and play by definition.
I would agree SpamSieve is relatively simple to set up and train for anyone who is even a little tech savvy and should be mostly smooth sailing after that. But my sense is it’s more of a power user tool. I don’t see “grandma” or anyone who usually asks for help with anything remotely technical even discovering, let alone installing, configuring and training a tool like SpamSieve. Fortunately, folks like us are often happy to do it for them.