[rrd-users] Can I get a single update into an RRA?

William Owen William.Owen at profilesmail.com
Fri Dec 11 18:41:29 CET 2009

I took a similar approach to Darren's suggestion to check for Terminal Services users connecting to a pool of servers. I log a 1 for connected, a 0 for disconnected and UN if offline. I cache the polling cycles to disk once a minute and use a script import the data every 15min which verifies that the data to be imported wouldn't result in a single entry. If the script detects a single entry will result, I do three updates, step-1:data/2, step:data, step+1,data/2. The graph isn't "accurate" but it will create a nice slope, which my management likes, and guarantee you get data. Calculating actual values is more accurate as you can use a CDEF to filter out anything other than 1 or 0.

-----Original Message-----
From: rrd-users-bounces at lists.oetiker.ch [mailto:rrd-users-bounces at lists.oetiker.ch] On Behalf Of A Darren Dunham
Sent: Friday, December 11, 2009 11:25 AM
To: rrd-users at lists.oetiker.ch
Subject: Re: [rrd-users] Can I get a single update into an RRA?

On Fri, Dec 11, 2009 at 09:42:11AM -0700, David Mitchell wrote:
> Optimally, it would be nice if there was some configuration using
> RRA:LAST or something which would result in the RRA getting populated
> with a single update. I thought about trying to fudge it by creating the
> RRD's with an extremely small step size (like 1 second) and always
> pushing two updates (N:5 -3:5 for example) at once to try and force data
> into the RRAs. It seems like there should be an easier way though.

I don't know of anything that is internal to RRD that would help you,
but you could fake it with an update wrapper.

Have the wrapper check the RRD for the last update.  If it's outside the
heartbeat, update the previous (stepwise) data point with a zero, then
update your current data point.  more work on your part, but a wrapper
could hide much of the complexity.


rrd-users mailing list
rrd-users at lists.oetiker.ch

More information about the rrd-users mailing list