[openstack-dev] ODP: [Monasca] Sporadic metrics
erickson at lsd.ufcg.edu.br
Wed May 18 16:45:47 UTC 2016
Yes, I'll be able to follow up your implementation in both monasca-ui and python-monascaclient. If I have any doubts, I'll not hesitate to contact you. Thanks for your support. :)
----- Mensagem original -----
De: "Tomasz Trebski" <Tomasz.Trebski at ts.fujitsu.com>
Para: "witold bedyk" <witold.bedyk at est.fujitsu.com>, "Erickson" <erickson at lsd.ufcg.edu.br>, openstack-dev at lists.openstack.org
Enviadas: Quarta-feira, 18 de maio de 2016 7:36:56
Assunto: ODP: [Monasca] Sporadic metrics
Erickson, do you know if you will be able to follow up my current implementation in either of components Witek has mentioned ?
If you would need some better background, let me know all your questions.
Currently the most crucial part would be the monasca-ui. This change would have to be created on top of the change for
non-deterministic alarms, because that way it only makes sense. The idea I have for it is to automatically propose type of an
alarm for the user based on the picked metric.
- non-deterministic for sporadic metrics
- determinist for sporadic metrics
Of course that is only proposition and user/operator can still make his own choices about that.
As for the python-monascaclient, most of the changes would be around metric commands.
I think there's actually nothing to be done about other commands (especially related to alarms definitions). But just
recently I've started using this CLI client, so I am no expert about that.
PS. Please let me know, if you have any doubts.
Od: Witek Bedyk <witold.bedyk at est.fujitsu.com>
Wysłane: 13 maja 2016 17:44
Do: Erickson; openstack-dev at lists.openstack.org
Temat: [Monasca] Sporadic metrics
thanks for your interest in our project and sorry for replying so late.
We are happy that you're willing to contribute to the development of
The incentive for this development is to introduce the possibility to
create alarms in Monasca based on event like metrics. E.g. A metric is
generated every time when an error in the given log-file appears. Such
metric is different from the standard Monasca metric in the way that it
does not appear on the regular, periodic basis. You can read a little
more about it here , but the blueprint itself is outdated and the
The main logic for alarm creation has been moved to
"deterministic_alarm" topic  and Tomasz works currently on that. Your
reviews are very welcome.
The changes in "sporadic_metric" topic  are meant to give the user
information that a given metric is of sporadic (non-periodic) character.
It will have no influence on monasca-thresh. Tomasz has put the code
that he already had in this topic but he must focus on
"deterministic_alarm" now, so please feel free to submit your patch-sets
if you want. Apart from the existing changes also changes in monasca-ui
and python-monascaclient repos are needed to display the 'sporadic'
attribute of the metric for the user. You could compare this change in
If you have questions to the implementation of "sporadic metrics", I
think the existing changes in gerrit  are the best place for exchange
and communication at the moment. Actually I should have updated the
blueprint for that. Sorry, my fault.
So, have fun with the code :)
P.S. Please note, I will be in holidays for the next two weeks.
FUJITSU Enabling Software Technology GmbH
Schwanthalerstr. 75a, 80336 München
Telefon: +49 89 360908-547
Telefax: +49 89 360908-8547
Sitz der Gesellschaft: München
AG München, HRB 143325
Geschäftsführer: Dr. Yuji Takada, Hans-Dieter Gatzka
More information about the OpenStack-dev