Cacti each minute???

[Quindor]

Great deal, got it working also. Manny thnx's!

Quindor

[cylam]

I have some concern about adding RRA....

Once we change the "step" in data template, we basically shortened the duration of each Primary Data Point (PDP).

Then, the previously defined RRA like "Yearly (1 Day Average)", which was told to consolidate 288 PDPs to form a 1 Day Average, can now consolidate 288 PDPs to form a 1/5 Day Average.

Also, as the number of rows (or Consolidated Data Point, in RRDTool's term), is kept unchanged, the "Yearly" RRA now just holds 1/5 year of data.

I guess once we change the "step" in data template (PDP), we should make a whole new set of RRAs instead of one single "Hourly" RRA.

After reading the manual of RRDTool, I guess we can just change the crontab to poll data every minute. If I have read correctly, I guess the RRDTool will average out the 5 data point gathered during the 5 minute interval, and store the averaged value into a PDP. In this sense, we will lost some precision as the "each minute" stat is averaged into "5 minute" stat. But at least we won't miss any data given by the remote agent.

Please kindly correct if I have misunderstood the RRDTool and Cacti.

[StefSybo]

The timespan of a graph is defined in seconds, so I suppose the 1 year graph would still contain 1 year of data, though the sampling rate might indeed be 5 times as much (1/5 day samples)

[jehul]

hi!

i followed exactly your steps!

i duplicated the data template "Interface - Traffic" to "Interface - Traffic (1 minutes)" with all changes in "rras", "step", "heartbeat"
now i created new graphs
after this i eddited all "data sources" to point to data template "interface- traffic (1minutes)"
and deleted the old .rrd files if they existed.

but now the rrd files where not regenerated.
so i created them with:

/usr/bin/rrdtool create \
/home/xxx/cacti/rra/xxx_traffic_in_vlan0.rrd \
--step 60 \
DS:traffic_out:COUNTER:600:0:100000000 \
DS:traffic_in:COUNTER:600:0:100000000 \
RRA:AVERAGE:0.5:1:500 \
RRA:AVERAGE:0.5:1:600 \
RRA:AVERAGE:0.5:6:700 \
RRA:AVERAGE:0.5:24:775 \
RRA:AVERAGE:0.5:288:797 \
RRA:MAX:0.5:1:500 \
RRA:MAX:0.5:1:600 \
RRA:MAX:0.5:6:700 \
RRA:MAX:0.5:24:775 \
RRA:MAX:0.5:288:797 \


but now the graphs don't show any data.
the .rrd's file modified time also didn't change.

the cron is running every minute, cause other graphs do work correct.


edit:
now i modified the original "Interface - Traffic" and created the graphs for it.
now everything works well.
but how about my old graphs, which are all running at 5 mins step?
will they function proberly or will i run in any problems with the old graphs?


thanks for your replys!

[claudio]

I'm not familiar with defining RRA's.

I don't understand, why I have to define a timespan of 14400 for a hourly(1min avg) graph.
Should the timespan not define the number of seconds for a graph? So, I would set timespan to 3600, or am I wrong??

For what exactly is the Row value for?

Regards,

Claudio

[claudio]

ok, now i understand.
i also used timespan of 14400, but this is a 4 hours graph,l not hourly graph.

and i created a whole new set of RRAs as cylam said, because i think he's right...

graphs working well.

but i still don't know, for what exactly the row value is for...

[steverodgers]

If I can take a few moments to clarify this and make sure everyone understands - it's not such a black art

* the number of data points tells you how many measurements - or runs of the poller - it takes to store a single point in each RRA.
so, for a poller executing once per minute, an daily with 1 minute avg would have 1 point. a weekly with 30 min avg would have 30 points etc.

* the number of rows corresponds to the number of datapoints for each archive that you want to store - so for a daily rra with 1 min avg, you'd have 60x24 = 1440 rows, for a weekly rra with 30 min avg, you'd have 2*24*7=336 rows.

* the timespan figure represents the number of seconds that each rra displays by default on the chart screens. so a daily would have 86400 seconds here etc.

[wizard]

To make this work going forward, it also helps to change the default value in table poller_item, field rrd_step. Change it to 60, from 300.

If you don't change this, new RRDs you create will still average your data into 5-minute chunks.

After changing it, execute this sql to update your poller_item table:

UPDATE poller_item SET rrd_step=60;


BTW, I am using the 0.8.6e version of cactid in my setup, so that when multiple poller interval support comes out, I will be ready

[sictc]

I think you have possibly 2 results. Either the databases get 5 times larger to contain enough information to have enough data for up to 1 year. Or the resolution will be distorted...

Besides, most network snmp data I personally get is collected in a 5 min average anyway. Since I don't have access to the source code of IOS of CatOS, I can not change the frequency to every minute iso five.

It may be more tricky that at first it would seem to change the frequency to 1 minute intervals stead 5 minutes.

[hunteralpha]

mysql --user=root

use cacti;

update data_template_data set rrd_step=60 where rrd_step=300;

update data_template_rrd set rrd_heartbeat=120 where rrd_heartbeat=600;

exit

[arkade]

Do the mysql statements have to be repeated each time you add a new source?

[Ahuber]

Negative. When you define the DS, you can define the steps, such as 300, but in your case (per minute) it would be 60.

[cbrmig]

You must also modify the cron job to be executed every minute... it MUST work

omg... forgot smthing:
In Data sources (templates): Step must be 60 and Heartbeat should be 120.

Hi..i'm sorry to ask, but where is "cron job" ?

Thanks.

[gandalf]

For unix: /etc/crontab, /etc/cron.d/cacti or "crontab -l -u <user>". For windows: I don't know
Reinhard

[ben_c]

Do the Cacti dev group have any plans to make this a more 'in built' function? It would be great for this to be more of a feature of Cacti without having to hack stuff up to get it working.