[rrd-users] How to prevent peaks?

[Simone Felici - Alpikom Spa]

Hello to all!

Since some weeks I use rrd-tool to monitor many different interfaces on a
router.
I create then only a graph with the sum of all interfaces to have an
overview of the total traffic. (CDEF....)
All works great except that sometimes in my graph I see some strange peaks
(example from 2 - Mbit/s to 100Mbits/s and back to normal traffic).

Here an example on how I create one of the many rrd files (one per
interface):

rrdtool create eth001.rrd \
--start `date +"%s"` DS:eth0_in:COUNTER:600:U:U DS:eth0_out:COUNTER:600:U:U
\
--step 300 \
RRA:AVERAGE:0.5:1:2016 RRA:AVERAGE:0.5:6:1500 RRA:AVERAGE:0.5:144:3650 \
RRA:MAX:0.5:1:2016 RRA:MAX:0.5:6:1500 RRA:MAX:0.5:144:3650 \
RRA:MIN:0.5:1:2016 RRA:MIN:0.5:6:1500 RRA:MIN:0.5:144:3650

Maybe set a max value should be an idea...



After this, every 5 minutes a script in crontab collect all informations
about the different interfaces (snmpget).


It's possible that rrdtool need informations every 5 minutes and if the
script takes more than 5 minutes it can cause a peak?

Is there a way to prevent such a kind of peaks or to correct them?
If I can determinate the right rrd file in which is contained the "error"
(peak) I'm not be able well to correct the error, because it's in a COUNTER
format. Any suggestion on how correct?


Thank's for advice!

Bye

Simon

--
Unsubscribe mailto:rrd-users-request at list.ee.ethz.ch?subject=unsubscribe
Help        mailto:rrd-users-request at list.ee.ethz.ch?subject=help
Archive     http://www.ee.ethz.ch/~slist/rrd-users
WebAdmin    http://www.ee.ethz.ch/~slist/lsg2.cgi