I figured I'd post some correlation stuff we are doing in order to encourage others to share some more correlation work. I know the contest is done, but hoping more to leverage the userbase to share the knowledge
We're interested in looking at Cisco Failover so we have an alert setup to fire based on 104001*, 103001, 103004. If any of these occur, we use Task Triage to output this to a team for investigation. Eventually a ticket is created for the firewall team. This is helpful because we can tabulate if pairs of ASAs are failing over to pinpoint problems to a cable, switchport problems, bad hardware etc based on trending and metrics.
Another we have is monitoring for devices not sending data. Since we're the log service, we really need to know if a device stops sending for some reason.
We've got it setup for our IPS that if we don't get in any event within 61 minutes, then we fire an alert.
We look for event ID 508100 (which is a package error) and variable * , with a filter on Count = 0 and Variable "Local Address laddr" in the correct Device Group. This is working for us, but we do have problems with certain devices that are less chatty. We're thinking of splitting into groups based on site or average events. Maybe have one group that alerts if no events come in for 3 hours, and one for 30 minutes or something like that, but this is the bulk of the idea....
Actually, the summer contest is complete, but we are running a new one now, so please post away! Winner gets $150 and runner-up gets $50 (AmEx gift cards).
Thanks for the great content!
Good stuff as were looking into something along those lines. Not to hijack this thread, but is there a convienent way to show from all of your monitored devices when the last log was received? ie..timestamp That might help for monitoring of devices not sending data.
Yeah I'm not sure on that, I'll give it some thought. I wonder if you could take the timestamp value and then setup some filter that says if the timestamp is older than 1 hour ago....
Right now our way is to look for a message not coming in, which it does the job, but I think grouping devices into different groups based on how chatty they are would be helpful.
I was trying to create corelation rule for device not sending the log in 24 hours .I have used CRL 23 rule default one but no luck ....any sugession ?