[rrd-users] Re: Calculating traffic

[Serge Maandag]

Well, if I'm correct:

Your step is 60 and your heartbeat is 600. Your hires RRA has a xff of
0.5.

So your graph will have a resolution of 1 value per 10 minutes tops
(600). Your RRD expects to be updated every minute (60). That makes 10
updates per value. Your DS will have the average value of those 10
updates. If you miss more than 5 of those updates (0.5x10), you're value
will be regarded unknown.

Since you're accounting traffic, you could update your RRA at a much
larger interval than once per minute. It saves on cpu and snmp traffic.
Choose the largest interval in which your counters do not overflow. try
half an hour for starters. If your updates don't fail often, you don't
need multiple updates per value (which is called CDP). If so, make the
step size equal to the heartbeat of your DS.

  rrdtool create port_xxx.rdd --step 1800 \
  DS:in:COUNTER:1800:U:U                 \
  DS:out:COUNTER:1800:U:U                \
  RRA:AVERAGE:0.5:1:100                 \
  <--etc.-->

Serge Maandag

-----Original Message-----
From: Iñaki Martínez [mailto:sistema at euskal-linux.org]
Sent: dinsdag 25 juni 2002 11:22
To: rrd-users at list.ee.ethz.ch
Cc: Alex van den Bogaerdt
Subject: [rrd-users] Re: Calculating traffic



Hi Alex van den Bogaerdt!!!

 First, thanks for answering.... 

> >  so, i look at the number...
> > 
> >  36159401
> >  36148921 lesser
> >  36220119
> > 
> >  Some times i've got confusing numbers....
> > 
> >  have i something wrong??????
> 
> How so?  You're calculating the number of bytes send during the
> *last*two*hours*.  If you slow down sending (a lower rate) then
> the total amount will be lower.
> 
> In extremis:  if you stop sending traffic at all and do the
calculation
> again after two hours, the total will become 0.  This is true because
> you didn't send any byte *during*the*last*two*hours*.

 Well.... i will start form the beginning....

 The rrd file......

  rrdtool create port_xxx.rdd --step 60 \
  DS:in:COUNTER:600:U:U                 \
  DS:out:COUNTER:600:U:U                \
  RRA:AVERAGE:0.5:1:600                 \
  RRA:AVERAGE:0.5:6:700                 \
  RRA:AVERAGE:0.5:24:775                \
  RRA:AVERAGE:0.5:288:3725              \
  RRA:MAX:0.5:1:1:600                   \
  RRA:MAX:0.5:6:700                     \
  RRA:MAX:0.5:24:775                    \
  RRA:MAX:0.5:288:3725;

 3725 = 5 years!!!!

 I get via snmp each minute the trafic from the switch port xxx.

 How can i do this better???

 I mean, the rrd file is saving each 5 minutes while i get data each
minute.
 I do not understand very well the structure of the rrd file.

 The graphic (new update):

  --start -2h
  --end now
  DEF:in=port_xxx.rdd:in:AVERAGE
  CDEF:tin=in,7200,*
  DEF:out=port_xxx.rdd:out:AVERAGE
  CDEF:tout=out,7200,*
  AREA:in#00FF00:"IN  "
  GPRINT:tin:AVERAGE:"B/s\: %9.0lf "
  GPRINT:in:MAX:"Max\: %9.0lf "
  GPRINT:in:AVERAGE:"Ave\: %6.0lf "
  GPRINT:in:LAST:"Cur\: %6.0lf\c"
  LINE1:out#0000FF:"Out "
  GPRINT:tout:AVERAGE:"B/s\: %9.0lf "
  GPRINT:out:MAX:"Max\: %9.0lf "
  GPRINT:out:AVERAGE:"Ave\: %6.0lf "
  GPRINT:out:LAST:"cur\: %6.0lf\c" >


 One thing i do not like is that IN (green) area is very rough, i think
that is problem of the rrd file that save each five minutes.


 Is it all OK?????

 Is there something that i can improve???


 Thanks in advance to all answers!!!!!!

--
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