Looking for CF explanation please.

Post general support questions here that do not specifically fall into the Linux or Windows categories.

Moderators: Developers, Moderators

Post Reply
thogan
Posts: 6
Joined: Thu Oct 06, 2005 9:14 am

Looking for CF explanation please.

Post by thogan »

So I finally got off my high-horse of if it isn't broken don't fix it and upgrade to 0.8.7d. In doing so several of my graphs stopped generating and after poking around the forums I discovered that I needed to change the graph templates from a CF type of “last” to “average” but only on the items that were use to plot the graph. Any item type of “gprint” could still use the “last” CF type.

So why is it that I cannot plot the “last” value recorded? It seems to me that the graph would not be an accurate representation of the data. Or do I not understand the difference between “last” and “average”?
alexs77
Posts: 41
Joined: Tue May 26, 2009 7:31 am
Location: Winterhtur, Switzerland

Re: Looking for CF explanation please.

Post by alexs77 »

thogan wrote: It seems to me that the graph would not be an accurate representation of the data.
Seems like, yes.

I'm plotting the amount of used licenses - see the attached graph.

As you can see there, at about "right now" (ie. right side of the graph), it looks as 2.8 licenses are used.

This cannot be. The script which counts the licenses always returns "full" numbers (ie. 0, 1, 2, 3, ...).

I'm using the SNMP - Generic OID Template. I understand that it plots the average value. But for "(snmp_oid): Current:" it prints "LAST" and in the attached screenshot, you can see that Current: 2.80. As I said, that's an impossiblity - the last value CANNOT be 2.80. It's either 2 or 3.
Or do I not understand the difference between “last” and “average”
For me, this would be a true statement :)
Attachments
graph_image.php.png
graph_image.php.png (26.83 KiB) Viewed 782 times
User avatar
gandalf
Developer
Posts: 22383
Joined: Thu Dec 02, 2004 2:46 am
Location: Muenster, Germany
Contact:

Post by gandalf »

There _may_ be a misunderstanding on rrdtool principles. RRDTool does _always_ perform some "nice" math to match the polling interval. This is the main reason for non-integer results being saved to rrd files even if the raw reading itself was an integer. You may want to consult rrdtool documentation
Reinhard
Post Reply

Who is online

Users browsing this forum: No registered users and 1 guest