0

I have some RSS feeds that are periodically fetched, but the time between each check is essentially random. After each fetch, the number of new items is logged along with the time of the fetch.

I want to make a graph using RRDTool which combines the data from each feed and shows the number of new items per hour over the course of a day.

For instance, I want it to show what the rate is for all feeds between the hours of 17:00 - 18:00, 18:00 - 19:00, etc. I want to use the unit new items/hour.

I have looked at the man pages and a tutorial, but the concepts of using multiple data sources and RRA's in this particular scenario is a bit over my head. I understand that I want to input the data using something like this:

rrdtool update feed1.rrd 1445113625:14 # 14 new items
rrdtool update feed2.rrd 1445113627:8  # 8 new items
rrdtool update feed3.rrd 1445113629:22 # 22 new items

rrdtool update feed1.rrd 1445113748:4  # 4 new items
rrdtool update feed2.rrd 1445113750:8  # 8 new items
rrdtool update feed3.rrd 1445113752:16 # 16 new items

How would I set up the rrdtool create and rrdtool graph commands to get the results I am looking for?

hololeap
  • 1,156
  • 13
  • 20

1 Answers1

0

You want to use the ABSOLUTE datasource type:

rrdtool create feed1.rrd --step 300 DS:feed:ABSOLUTE:600:0:U RRA:AVERAGE:1:1000

this assumes you fetch your feed at least every 600 seconds

once you did a fetch, you run

rrdtool update feed1.rrd N:$number_of_articles_fetched

to create graphs, you load data from all rrds

rrdtool graph feed.png \
DEF:f1=feed1.rrd:feed:AVERAGE \
DEF:f2=feed2.rrd:feed:AVERAGE \
CDEF:sum=f1,f2,+ \
AREA:sum#00f:Articles
Tobi Oetiker
  • 5,167
  • 2
  • 17
  • 23