Nagios Support Forum

Hi Team,

Please advise what is the duration before the rrd's are averaged?

My client wants the accurate data for 3 months and post that he is okay with averaging the rrd's. Please advise how we can change this.

Alright. We just went over this with someone from either you or your clients organization. This is possible, but you will lose all previous historical data. Additionally, the file we are going to change is the template for rrds and it may be overwritten with major xi upgrades, so beware of that.

The file you need to edit is: If most of your checks are on 5 minute intervals, edit the file so that is resembles: Essentially, we altered the 5 minute retention from 10 days (5min * 2880) to 90 days (5min * 25920). In order for these changes to take effect, you will need to remove/move all your rrds from /usr/local/nagios/share/perfdata/<host>/.

Let me know if you need more granularity than 5minutes.

When we do a rrd fetch we get a csv with the number of the datasource at the top of the sheet, Is it possible to get the name of the datasource as we see in the graph instead of a number?

Well, you would have to alter the php script for fetching rrds, but it may lead to issues as the javascript looks for datasources by number, so we would need to change a few other php scripts. . . So, not really?

Hi Team,

We are planning to increase the RRD rentention today, Once we move/remove the current rrd's from the location /usr/local/nagios/share/perfdata/<host>/ would a new folder be created automatically? Also, Can we remove/move the entire perfdata Folder, so that all the servers generate the required data.

You can move all the old RRD's out of it, do not delete the folder itself, they should be generated when performance data is returned to Nagios. It may take a few checks though, before you see it displayed.

I have made some changes to the vlaues, we want to store 1 min data for 1 week, 5mins for 90 days, 30mins for 6 omnths and 360mins for a yr.


PNP default RRA config
#
# you will get 400kb of data per datasource
#
# 2880 entries with 1 minute step = 7 days
#
RRA:AVERAGE:0.5:1:10080
#
# 2880 entries with 5 minute step = 90 days
#
RRA:AVERAGE:0.5:5:25920
#
# 4320 entries with 30 minute step = 180 days
#
RRA:AVERAGE:0.5:30:8640
#
# 5840 entries with 360 minute step = 356 days
#
RRA:AVERAGE:0.5:360:1460

RRA:MAX:0.5:1:10080
RRA:MAX:0.5:5:25920
RRA:MAX:0.5:30:8640
RRA:MAX:0.5:360:1460

RRA:MIN:0.5:1:10080
RRA:MIN:0.5:5:25920
RRA:MIN:0.5:30:8640
RRA:MIN:0.5:360:1460


Are the values OK?

They look fine. Just understand that these settings will lead to *significantly* larger RRDs.

We have set aside a large partition for the RRD's as the client has asked for this Data, but I would like to understand the strain on RAM for these settings, Would this change lead to high memory(RAM) utilisation?

These changes will lead to higher *everything*. Any processing of the RRDs will take more memory as the RRDs will now be substantially bigger. They will also take more disk space for the same reason. When values are added, cpu usage will probably be a bit higher as well as there will be many more steps to compress/aggregate.