Skip to main content
rict
New Member
August 27, 2015
Question

Typhoeus filter?

  • August 27, 2015
  • 5 replies
  • 8190 views

We have a collection of web servers that are getting "blasted" by traffic by Typhoeus spiders. When I say "blasted", I mean over 1,500 requests/minute. The web admin was able to get the IP addresses for the sources, but they are all Amazon AWS addresses, which means we're only seeing the Amazon address, not the root/source address.

 

Of course, Amazon is no help, since they respond that "Amazon EC2 Public IP addresses may change ownership frequently, without additional information we will be unable to identify the correct owner of the IP address for the period of time in question". We've given them all kinds of log info, and modified the Apache servers to reject these requests, but we'd like to create a filter of some kind on the firewall itself.

 

Here's an example of some info from the request log:

"GET /en/util/conflict-of-interest.html HTTP/1.1" 403 - "-" "Typhoeus - https://github.com/typhoeus/typhoeus" 227 -

 

I'm not a FortiGate expert of any kind...does anyone know if there's current Typhoeus filter of some kind available, or failing that, if a filter for this can be created?

 

rict

    5 replies

    xinger
    New Member
    August 28, 2015

    I'm no expert either.  Here's a manual for do-it-yourself...

    http://video.fortinet.com/uploads/documents/IPS%20Signature%20Syntax%20Guide.pdf

     

    Or options for having Fortinet to do it for you...

    http://www.fortiguard.com/more/fortiguardpremier

     

    I don't have experience with either way.  Good luck.

     

    emnoc
    New Member
    August 28, 2015

    I have to agreed , writing a simple  IPS filter with a ban action and expiration is what I would do.

     

    Here's a few examples;

     

    http://socpuppet.blogspot...es-fortinet-style.html

     

    http://docs-legacy.fortinet.com/fos50hlp/50/index.html#page/FortiOS%205.0%20Help/ips_signatures.153.44.html

     

    http://socpuppet.blogspot.nl/2014/08/how-to-write-ips-signature-to-block.html

     

    Since the sources are many and from  AWS, you could maybe narrowed it down to those ranges but it's interesting as to why your webservices are being hit and from only AWS  addresses?

     

     

    gschmitt
    New Member
    August 28, 2015

    rict wrote:

    "Amazon EC2 Public IP addresses may change ownership frequently, without additional information we will be unable to identify the correct owner of the IP address for the period of time in question".

    WTF? Is that even legal?

    emnoc
    New Member
    August 28, 2015

    And what would you do differently if you knew who the owner was ?

     

    Regardless of the owner information, he has a small HTTP_GET flood taking place. He's best bet is to mitigate b4 nailing up a front-end DDoS Proxy-mitigator serviecs & to write a IPS rule that Bans and BL the src_ipv4 address for a controlled amount of time and hope the attack sources that could be just infected dies down and go away.

     

    ken

    ede_pfau
    SuperUser
    SuperUser
    August 28, 2015

    Why not keep it simple and create a flood DoS policy? Easy if you're using v5.2.x. Type would be "TCP flood" which is not very specific, i.e. not HTTP specific. But at that rate you would catch that source IP anyway. Then action=block, not even quarantine.

     

    Regarding Fortinet support, we've had posts here in the forums where Fortinet support wrote an ad-hoc IPS filter for a customer. As I understood at that time, within the limits of a support case. Would be worth a try if the above doesn't work for you.