I use snmp quite heavily in monitoring our 13000 endpoint lan. I measure the
load by looking at the in/out octets on my monitoring systems. I read the
octets, packets, errors, bcast, mcast and discards out of each endpoint port
(switch or router) every 5 minutes. These activites seem to put little load on
the network support devices. I also monitor the cpu utilization of the devices,
and there is no noticable difference between a day with out snmp monitoring and
a day with it.
Of course the "load" also depends on how much processing the device being
accessed has to do to return the response.
"Higher" load inquiries are reading bridgetables and arp cache dumps, however
there is no reason to do those "frequently". I do it twice a day to model the
network.
Hope this helps....
Connie Logg, Network Analyst, [EMAIL PROTECTED], 650-926-2879
Stanford Linear Accelerator Center, MS 97, 2575 Sand Hill Road, Menlo
Park, CA 94024
"'IF' is the middle word in life" but "Happiness is found along the way, not at
the end of the road."
-----Original Message-----
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]
Sent: Tuesday, June 17, 2003 1:12 PM
To: [email protected]
Subject: [rrd-users] rrdtool - graphing the load of snmp itself
When one becomes successful with RRD and mrtg, there are times that the
periodic question comes up: "how much snmp load is this putting on my
network"?
I get this with all my snmp pollers. Especially a few weeks back when one
of them went nuts and looped while doing discovery of the media2ip table.
Ended up driving the IPSNMP process CPU on the cisco switch very high. So... I
am looking to track the amount of SNMP traffic by my snmp pollers.
They are on a mix of windows and unix. products like mrtg, concord,
netview, spectrum, etc...
1) I was going to graph the mib2 snmp group - something like
snmpInGetRequests. But none of my snmp agents on these servers return any
values for the snmp group of mib2. sigh...
2) I have been looking instead at the icmp group , like icmpInMsgs . these
seem to have values. but it is at all meaningful? my busiest poller
seems to be at around 4/second.
not great solutions. so, a question:
When asked to track over time the snmp load of your pollers, what do you
use?
much thanks for any help
Don Mahler
SAIC /Telcordia
--
Unsubscribe mailto:[EMAIL PROTECTED]
Help mailto:[EMAIL PROTECTED]
Archive http://www.ee.ethz.ch/~slist/rrd-users
WebAdmin http://www.ee.ethz.ch/~slist/lsg2.cgi
--
Unsubscribe mailto:[EMAIL PROTECTED]
Help mailto:[EMAIL PROTECTED]
Archive http://www.ee.ethz.ch/~slist/rrd-users
WebAdmin http://www.ee.ethz.ch/~slist/lsg2.cgi
