I'm looking for some sort of bandwidth meter to watch our dedicated server customers.
I know there are so many out there for individual servers, bandmin, vstats and iftop to name a few, but I'd like to find one that we can put on a server, but we can have it setup to log the bandwidth on a network switch or a Cisco router? As for Cisco, we have two main routers on our network a 3620 and a 2100 series.
I think we might have on part of our network a switch that can monitor incoming and outgoing per port, but it cannot be logged into remotely.
I'm looking for graphs when customers request them, excel sheets for customers, etc., etc. I've seen other companies provide these.
Any suggestions or help would be greatly appreciated.
That HotSaNIC looks like something to be installed on individual servers, is it not?
I want to monitor our dedicated server customers without installing anything on their servers in specific.
Ah, sorry, I misread your post there.
Yes, HotSaNIC is a tool for individual machines primarily. I believe it can be used to chart results from other machines (or routers) obtained via SNMP or similar but not without some jiggery-pokery you can probably be doing without!
That is a bad idea ClineCOM.
If I am not mistaken, traffic such as DDOS that can be filterd at the firewall or any traffic that hits iptables and gets denied on the machine will not be graphed.
SNMP does not hurt really hurt the switch performance and you are not installing anything on the switch.
You install it on one machine which sends snmp requests to the switch which tells it how much data is going through each port.
All you do is configure a simple parameter on the switch (SNMP Community String)
Cline any SNMP graphing program will work, I suggest you look at RTG/MRTG/Cacti/HotSanic demo graphs and decide on which to use.
Some will be straight forward on setup, while others will be more difficult as usual.
But sometimes the best solutions are not the easiest. GL
What do most of you hosts use to monitor webhosting clients on your servers individually? I'm curious as to how to set limits, and when those limits are reached, automatically disable that client, in linux of course.