[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: draft-ietf-opsec-logging-caps-03




On Jul 3, 2007, at 12:55 PM, Roland Dobbins wrote:


On Jul 3, 2007, at 11:39 PM, Warren Kumari wrote:

One possible concern with rate-limiting syslog / SNMP / whatever is that people will then try and generate non-critical events that cause alerts to be generated in the hope that they can then slip a more nefarious event in and have it also be caught by the rate- limiter.

Possible, but hard to accomplish without inside knowledge (which is certainly a possibility, of course).

Yup -- very true, although I did see it happen once -- It was a failed attempt to brute force TACACS. There were ACLs configured on a device with the "logging" keyword that caused the device to log all ACLs hits and a (not so bright) employee that had read-only access to the syslog server -- he noticed that if he did "ping -f" (which he could easily explain as a diagnostic process that had gone awry or something) to a device behind the ACL the device would be so busy logging ACL hits that it wouldn't log failed login attempts. What he *didn't* check first (I did mention that he wasn't so bright) was whether failed logins usually showed up in this syslog server.....

I had configured syslog-ng to only forward a very small subset of logs to the server that he was looking at (the logs were for use by a NOC tool that watched the logs for interface flaps, ACL hits and one or two other similar issues)...

The TACACS server logged somewhere between 8,000 - 9,000 failed logins before I reacted to the alerts that it was generating, but looking through the aggregated syslog I only saw around 2,000 attempts, so I guess that his trick did work to some extent :-)


What would be great would be, if there were a rate-limiter it would prioritize high severity messages over lower severity ones (whee! QoS for logs!).

How about multiple rate-limiters for each individual log-type, as well as for each severity/informational level?

That would be be work, although I am concerned that the overhead might overtax the device's processor...

My main comment is that, for local logs, the oldest, least important messages are dropped before old important messages....

W



---------------------------------------------------------------------- -
Roland Dobbins <rdobbins@cisco.com> // 408.527.6376 voice

       Culture eats strategy for breakfast.

               -- Ford Motor Company





--
Curse the dark, or light a match. You decide, it's your dark.
                -- Valdis Kletnieks