Question

I want to extract some hourly data from rrdtool databases in order to create some graphs within a dashboard system.

These databases don’t have an hourly datasource, the closest is a 30-min datasource (they are generated by munin)

Now, I can use rrdfetch, but that doesn’t do the nice averaging that rrdgraph would do, so something like this

rrdtool fetch xxx-apache_accesses-accesses80-d.rrd AVERAGE \
   --resolution 3600 -s 1328458200 -e 1328544600

Might give me 30 min data points like this

2012-Feb-05 16:30:00 3.5376357135e+00
2012-Feb-05 17:00:00 3.4655067194e+00
2012-Feb-05 17:30:00 4.0483210375e+00
2012-Feb-05 18:00:00 4.3210061422e+00
....

I could average those, but it seems that rrdgraph can output parsable text, but I can’t figure out the correct incantation. Here’s what I’ve tried

 rrdtool graph dummy.png -s 1328523300 -e 1328609700 \
     DEF:access=xxx-apache_accesses-accesses80-d.rrd:42:AVERAGE \
     "PRINT:access:AVERAGE: %5.1lf %S"

outputs

0x0
   4.7

Now I think that’s simply the average for the period given, but is there any way to get rrdtool to spit out an average for particular chunks or step sizes? I tried --step but this did not change the output.

I could call rrdtool graph for each data point I need, but that seems rather wasteful.

Answer 1

No sooner had I posted than I hit upon the right approach!

rrdtool xport -s 1328523300 -e 1328609700 --step 3600 \
    DEF:access=xxx-apache_accesses-accesses80-d.rrd:42:AVERAGE \
    XPORT:access:"average"

This gives me the dump I need…

<?xml version="1.0" encoding="ISO-8859-1"?>

<xport>
  <meta>
    <start>1328526000</start>
    <step>3600</step>
    <end>1328612400</end>
    <rows>25</rows>
    <columns>1</columns>
    <legend>
      <entry>average</entry>
    </legend>
  </meta>
  <data>
    <row><t>1328526000</t><v>2.1949556516e+00</v></row>
    <row><t>1328529600</t><v>2.0074586816e+00</v></row>
    <row><t>1328533200</t><v>2.4574720485e+00</v></row>
    <row><t>1328536800</t><v>3.4861890250e+00</v></row>
    <row><t>1328540400</t><v>4.2725023347e+00</v></row>
    <row><t>1328544000</t><v>6.2119548259e+00</v></row>
    <row><t>1328547600</t><v>5.6709432075e+00</v></row>
    <row><t>1328551200</t><v>6.1214185470e+00</v></row>
    <row><t>1328554800</t><v>8.1137357347e+00</v></row>
    <row><t>1328558400</t><v>5.8345894022e+00</v></row>
    <row><t>1328562000</t><v>6.2264732776e+00</v></row>
    <row><t>1328565600</t><v>6.1652113350e+00</v></row>
    <row><t>1328569200</t><v>5.8851025574e+00</v></row>
    <row><t>1328572800</t><v>5.4612112119e+00</v></row>
    <row><t>1328576400</t><v>6.3908056120e+00</v></row>
    <row><t>1328580000</t><v>6.0361776174e+00</v></row>
    <row><t>1328583600</t><v>6.3164590113e+00</v></row>
    <row><t>1328587200</t><v>6.0902986521e+00</v></row>
    <row><t>1328590800</t><v>4.6756445168e+00</v></row>
    <row><t>1328594400</t><v>3.9461916905e+00</v></row>
    <row><t>1328598000</t><v>2.9449490046e+00</v></row>
    <row><t>1328601600</t><v>2.4011760751e+00</v></row>
    <row><t>1328605200</t><v>2.2187817639e+00</v></row>
    <row><t>1328608800</t><v>2.1775208736e+00</v></row>
    <row><t>1328612400</t><v>NaN</v></row>
  </data>
</xport>