Spam overload: What countries?

I've posted a couple of times before about the Mollom service I use to filter comments and e-mails from this site for potential spam.

Lately, the spam on this site has gone just crazy. Happily, Mollom has caught it all - I've not had one get through.

In case you can't see the numbers on the graph, here's how it's gone:

• 7th November = 337
• 8th November = 261
• 9th November = 407
• 10th November = 242
• 11th November = 299
• 12th November = 742
• 13th November = 1121
• 14th November = 1457
• 15th November = 2379
• 16th November = 2801
• 17th November = 1288
• 18th November = 761
• 19th November = 791
• 20th November = 1262

That's 2801 spam comments per day at its peak. One every 30 seconds! In the old days, I used to rely on occasionally checking over the blog to see if anyone had left a spam comment. Then I installed a tweak that e-mails me if a comment is left so I can pick it up a bit quicker. Then I set Mollom to work, so that the spam gets filtered at source. If I had to remove 2800 comments each day using a manual process, I'd do nothing else!

So where do they all come from?

Well, what follows will be an approximation, as I've looked at the data for the past 4 days, and only looked into those computers (IP addresses) that spammed me 4 or more times. Between them, those IP addresses account for 3,157 spam attempts.

In order of prolificity, here is where those spammers were from:

• China: 2708
• Russia: 320
• United States: 64
• Ukraine: 24
• Poland: 22
• Canada: 8
• Sweden: 7
• Germany: 4

So there you have it. China and Russia between them account for 3,028 of those spam posts, which is 95.9% of the spam.

Mollom continues merrily stopping my spam. But I'm sorry if you visit me from China or Russia as a real human visitor. I'm going to block those two countries from Apache as way of trying to get this torrent of potential spam under control.