OK, I suppose, but I was thinking the issue was "how do you alert on a point in time graph reading?" and my take was "don't, alert on a recent trend reading". But I definitely appreciate that a change in graphing backend would be monumental!

Thanks,
Mark

On 3/1/2015 12:57 PM, Adam Armstrong wrote:

Oh, and it's utterly irrelevant to any question about alerting.

adam.

On 02/03/2015 02:57:39, Adam Armstrong <adama@memetic.org> wrote:
I looked at using Graphite/Ceres a couple of years ago, but I couldn't get graphs out of it with good enough layout so I gave up.

It would be a colossal amount of work, and at this point, I don't think it would bring benefits greater than other things we would spend our time on.

adam.

On 02/03/2015 02:52:28, Mark D. Nagel <mnagel@willingminds.com> wrote:
On 2/28/2015 8:23 AM, Adam Armstrong wrote:

We don't have the ability to alert on metrics polled as 'graphs' yet. I'm still trying to work out how best to do it.

I would think something along the lines of Graphite's queries for moving averages over a timespan would be a good start. Which of course begs the question -- does it make any sense to use Graphite as the backend instead of RRD? Huge change, granted, but it is much more scalable among other nice attributes. More machinery, though. Without actually using Graphite, would it be even reasonable to add a checker type linked to a RRD query? I am guessing it would be too much load past a handful of checks.

Regards,
Mark

This message has been scanned by CanIt-PRO.

_______________________________________________ observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium

-- 
Mark D. Nagel, CCIE #3177 Emeritus <mnagel@willingminds.com>
Principal Consultant, Willing Minds LLC (http://www.willingminds.com)
cell: 949-279-5817, desk: 714-495-4001, fax: 714-646-8277

** For faster support response time, please
** email support@willingminds.com or call 714-495-4000