[rrd-users] Re: Blocky graphs
Alex van den Bogaerdt
alex at ergens.op.het.net
Mon Sep 13 10:21:08 MEST 2004
On Mon, Sep 13, 2004 at 08:55:07AM +0200, Stefan Coetzee wrote:
> I did read the tutorial, but the examples are for very short periods of
> time, but thanks all for the help,
So? Examples are just that: examples. By writing a tutorial I hope to
share knowledge. If you need two years worth of data instead of one year,
apply the same logic and reasoning and come up with your own script.
> I'll try see where I can fix this, I know there's something small and stupid
> I'm doing wrong tho.
Start with an image with a known size (say: 400 pixels). Start by
displaying one hour on each pixel-column, thus 400 hours. Make sure
your database contains >>>at least<<< 400 hours worth of data in an
RRA with a resolution of one hour per row. Alternatively, keep at
least 400 hours worth of data in an RRA with a resolution of 30 minutes
per row (thus: more than 800 rows). Another alternative would be to
keep > 100 rows of 4-hour data. This would start to look "blocky" as it
would show 100 rows each 4 pixels wide.
In all cases, there's (more than) 400 hours available.
> What I've tried, that does work, but is not efficient, is keeping a year's
> worth of hires data,
> or 5minute averages..
And why didn't you try keeping a year's worth of 2-hour data ? That would
be much more efficient (4392 rows instead of 105408).
But if you want to plot "--end oct 14 00:00 --start end-12h" and if
you want that "not blocky", you have no choice but to keep the data
in high-resolution. After all, RRDtool cannot use what it cannot fetch.
You want an answer? You'd better follow the following guidelines!
Linesize well below 80 chars. Reply to the list, not to me. Trim
irrelevant lines. Answer below the remaining paragraph.
Unsubscribe mailto:rrd-users-request at list.ee.ethz.ch?subject=unsubscribe
Help mailto:rrd-users-request at list.ee.ethz.ch?subject=help
More information about the rrd-users