XF 2.2 Is there a performance Impact by adding 5k Spam phrases with wildcards?

[Fred.]

Hi,

I have several add-ons to prevent spam and akismet, stopforumspam and all you can get...
But I still get spam. And these days it's mostly Chinese propaganda spam.

So... I got a list with 5000 of the most used Chinese characters and added * before and after to make it a wildcard.
So it looks like this.

*的*
*是*
*不*

Is there going to be a noticeable impact in performance if I put 5000 lines in the spam phrases?
I think it might slow down a lot. But I have no other good options.

 

[Paul B]

You can add them to a single entry using the regex or condition - /(的|是|不)/ui

Have you considered using iptables at the server level to block specific country IPs (assuming you don't have legitimate visitors from there).

 

[Fred.]

You can add them to a single entry using the regex or condition - /(的|是|不)/ui

Have you considered using iptables at the server level to block specific country IPs (assuming you don't have legitimate visitors from there).

Thanks Brogan.
So so 5k characters in a single regex would preform better?

I'm using iptables, but they use VPN's or botnets.
No, my forum is in Dutch so no legimate Chinese speaking visitors. But I try not to block whole countries because it's a travel forum. So there might be Dutch people visiting my forum in China. But that chance is probably small.

 

[Paul B]

So so 5k characters in a single regex would preform better?

Hard to say - 5,000 is a lot of entries.

You may have to do some testing to see how each one performs.

Is it not possible to pick out the most common characters used in the most recent spam and just stick to those?

 

[Pepe2012]

Hi,

I have several add-ons to prevent spam and akismet, stopforumspam and all you can get...
But I still get spam. And these days it's mostly Chinese propaganda spam.

So... I got a list with 5000 of the most used Chinese characters and added * before and after to make it a wildcard.
So it looks like this.

*的*
*是*
*不*

Is there going to be a noticeable impact in performance if I put 5000 lines in the spam phrases?
I think it might slow down a lot. But I have no other good options.

I limit all new users to three posts to catch spammers, and their first ten need mod approval is case it's a previously banned member trying to sneak back in.

Oh, and I use a captcha question.

 

[Masetrix]

OT?

Hello,

you will only load the server unnecessarily with such queries. This even offers the possibility for DDOS ...

We "only" use StopForumSpam to fight spammers.
New users are only allowed to post in a few defined areas, must have created at least 15 postings and be registered for at least 15 days before they are moved to a user group, where they have further options.

I think 15 days are enough to realize what someone really wants. If you also forbid the use of a signature, conversations and profile postings during these 15 days, you are already well on the safe side.

 

[briansol]

block at the firewall level with cloudflare.
pose a challenge if they are real instead

handling this in front end software is the wrong layer IMO

 

[Chromaniac]

block registration page access through cloudflare instead. sure vpns would still bypass it, but it has massively improved spam situation on my own forum.

 

[Fred.]

Apparently the Chinese alphabet has 50000 characters so that's why there are still so many... I might be able to do it with less.
Ok, So the solution doesn't seems to be the best. So I need to look at other options. Good that I asked here.

I'm the only mod on the forum and I'm really busy with my job at the moment so I try to automate as much as I can.
Manual approval only frustrates real users... If you have mods online all the time this might not be a problem. But in my case it is.

I will look into cloudflare.
Is this possible on the free plan? or do I need a paid plan for this?

 

[Chromaniac]

it can be done on free plan. five rules are free.