How to block a spam attack?

Hi,

Since this morning, my Discourse is under a spam attack. They create a new account and post 5 messages.
I delete the spammer each time, but since they continue to create new accounts, I don’t know what to do.

Here the spams, maybe some of you get the same:

Thank you for your help

1 Like

Do you have askimet installed/configured?

3 Likes

No, how can I do that?

Edit: I found the discourse-askimet plugin. Should I install this version?

1 Like

That’s the one unless there’s another version somewhere… it’s pretty straightforward from what I saw- as our site isn’t live yet I can’t attest to its effectiveness, but others I’ve spoken to have had nothing but praise for it.

Let me know how it goes - We’ve been hit by a Bamwar (on our old forum) in the past and it gets quite annoying to clean up after…

Ok thanks,
I found no mention of this plugin on this forum. Can someone give me some feedback on it?

Akismet is the reference when it comes to spam. It takes a bit to learn about the content of your site, but once it’s rolling, it’s awesome.

Do you see much spam on meta? :wink:

2 Likes

To be clear, these are human beings, so they will easily bypass any Capcha. You need the official Akismet plugin to block them based on Bayesian content filters.

(Can also block by IP, but they have a lot of IPs…)

1 Like

I just installed the Akismet plugin and it already blocked the first spam :smile:
Thanks !

PS: I’m translating the plugin into French

8 Likes

The Korean bamwar human spammers are still around and still pernicious enough to be effective even with fast typist and Akismet checking and new user post limits and IP limits. @sam will add language specific site settings (we need this for all uppercase checks and other stuff anyway) to make sure the common thread of Korean text is a huge red flag for a new user post.

4 Likes

This is something that comes up a fair bit, I am adding the feature to the 1.8 list

3 Likes

This topic was automatically closed after 3244 days. New replies are no longer allowed.