[rrd-developers] rrdupates per second

Scott Brumbaugh scottb at prolexic.com
Tue Jun 24 17:11:59 CEST 2008

Hi Florian,

On Sat, Jun 21, 2008 at 06:54:33PM +0200, Florian Forster wrote:
> Hi Scott,

To be clear, this quote was from Ian Holsman in a previous mail.

> > > my thinking was that we could have a RRD Daemon running that would
> > > the CLI would talk to instead of accessing the files directly.
> I've had the same idea, but never really had the _need_ to actually sit
> down and write such a daemon. You mail motivated me enough to finally do
> it, though ;)

> You can pull my Git repository from
>   <git://git.verplant.org/rrdd.git>
> or browse the source and change sets online at
>   <http://git.verplant.org/?p=rrdd.git;a=summary>

Thanks for the update it looks pretty cool and I appreciate the info.
Here is a short note outlining the features that our project was
looking for in a time-series store.  We found by adding a network
layer to between the rrd data fetch and graph we were able to satisfy
these.  I think our items are pretty general and other projects would
maybe make a similar list of requirements.

. Scalability, we want to make sure that as we add more metrics we
don't hit a performance limit.

. Distributed, we want to collect and aggregate metrics from
geographically dispersed locations and be to display them with a
centralized application.

. Reliability, we want the system as a whole to continue working even
if a portion of our distributed database was unavailable.

. Performance, We need the database to populate quickly but we also
want to read out the latest updated rates quickly.

. Cost, Of course low here, both in cash but also in maintenance
overhead and complexity.

> Florian octo Forster
> Hacker in training
> GnuPG: 0x91523C3D
> http://verplant.org/


Scott B

More information about the rrd-developers mailing list