To answer your direct request, it looks like the RRD Create command will
error unless it includes a 1 step AVERAGE statement:
RRA:AVERAGE:0.5:1:525600
RRA:MAX:0.5:4:525600
Ultimately, I'm not sure that Zenoss Gui zoom function is best suited to
presenting the precision that you want. Any time you scale data to fit a
specific size (in any application), averaging data will be an inherent
part the process of building the graphs. Quoting the RRDgraph manpage:
The data fetched from the RRA is then consolidated so that there
is exactly one datapoint per pixel in the graph.
Further elaboration is provided in the rrdgraph_data manpage.
If the resolution of the data is higher than the resolution of
the graph, the data will be further consolidated. This may
result in a graph that spans slightly more time than requested.
Ideally each point in the graph should correspond with one CDP
from an RRA. For instance, if your RRD has an RRA with a
resolution of 1800 seconds per CDP, you should create an image
with width 400 and time span 400*1800 seconds (use appropriate
start and end times, such as --start end-8days8hours).
Ultimately, I don't think that the Zenoss zoom interface gives you the
control over the DEF statement that you require to generate a best
effort (and best effort is the best that RRDgraph will allow for this
use). You have two ways to obtain this level of accuracy with only one
RRA statement that I can see.
Option 1: You would need to either add 12 pixels per hour to the graph
width each time you zoomed out, eliminating the need to scale the data.
Option 2: To keep the same graph width use a DEF statement that
specified a calculated date range that corresponds to a specified step
time.
DEF:checkWebsite1_time=/opt/zenoss/perf/Devices/webserver/checkWebsite1_time.rrd:ds0:MAX:step=300:end=$((`date
+%s`/300*300)):start=e-60d:reduce=MAX
The way most RRD based applications I have used (including Zenoss)
mitigate this problem is by using set time frames (I.E. Zenoss and
Cricket allow you to choose Hourly, Daily Weekly, Monthly, Yearly
graphs) with multiple RRA statements that correspond to those
resolutions. The only main difference I see in most other applications
is that they tend to consolidate data at the 30 minute mark instead of
the 20 minute mark.
I think that if you are really looking for determining the actual max
for a given time period, (I.E. the max value in an hour) a tabular
report would actually be a better tool to present the information you
want, not a graph.
On Mon, 2007-07-16 at 20:53 +0000, agthurber wrote:
> Please excuse me for feeling dense. I am understanding the concept but if you
> could provide an example think that would clear up the rest of the confusion
> I am having.
>
> I created a new data source, and inside that source i created a new data
> point called time and in the create cmd section of that data point i entered
>
> RRA:MAX:0.5:1:525600
>
> hoping that would collect data in the pattern i want. with 1 step per cycle
> so none of my data gets averaged while collecting. However the graph appears
> broken and shows no data. Could you please give me an example syntax to enter
> in to the create cmd section of a new data point so that it collects the data
> exactly as it comes in from my command running once every 5 minutes and so
> that it will not average out my data when zooming out on my graphs? This has
> been quite a complex task for something I would think ought to be an option
> in the zproperties!
>
> thanks very much for your help!
>
> ------------------------
> A. G. Thurber
>
>
>
>
> -------------------- m2f --------------------
>
> Read this topic online here:
> http://community.zenoss.com/forums/viewtopic.php?p=8776#8776
>
> -------------------- m2f --------------------
>
>
>
> _______________________________________________
> zenoss-users mailing list
> [email protected]
> http://lists.zenoss.org/mailman/listinfo/zenoss-users
--
James D. Roman
IT Network Administration
Terranet Inc.On contract to:
Science Systems and Applications, Inc.
_______________________________________________
zenoss-users mailing list
[email protected]
http://lists.zenoss.org/mailman/listinfo/zenoss-users
