Anope IRC Services

Anope Development => Feature Requests => Topic started by: ben on July 07, 2007, 10:45:02 PM

Title: services flood notice
Post by: ben on July 07, 2007, 10:45:02 PM
ok i know i havent looked really hard but to my knowledge it doesnt appear to be a module that alerts opers or services opers that anyone is flooding the services (besides OperServ, but thats only because the flooder has not access)

so if the flooder is smart... you wouldn't know your services are even being flooded

this includes annoying scripts that people have no clue how to use
(where the user gets akicked and he has auto-cs_unban script, he just does it over and over again)

so if there is one point me in the right direction please, if not well then this is a request
Post by: ben on July 25, 2007, 08:32:40 PM
Can someone at least leave a comment? :[
Post by: katsklaw on July 25, 2007, 09:23:19 PM
on IRC text flooding is a moot point because nearly all IRCd's (those that don't are old and should be upgraded anyway), support wan'ts called sendQ/recvQ, these to handy dandy features  have been around for about 10 years now and will automaticly kill users for sending too much text too fast, thus making it nearly impossible to flood services to start with.

In addition to that, clients should never be able to send so much information to services from a single connection to cause any form of latency to services. Nearly all clients would be on some form of asymetric connection so their upstream will max out before Services downstream will. If you have a single user that can, then take services off it's dial-up connection and put it in some real bandwidth.

Currently the only way to really flood services on IRC is with multiple clients, in which case it's not possible to tell a flooder with several clients from serveral really active clients sending acceptable traffic.

[Edited on 25-7-2007 by katsklaw]
Post by: fudge on October 01, 2007, 06:42:07 AM
Besides operserv would only be reporting the flood, not the pseudo client actually in question.