[Openstack] Standardizing External APIs

Monsyne Dragon mdragon at RACKSPACE.COM
Wed Sep 7 19:55:47 UTC 2011


On Sep 7, 2011, at 1:19 PM, Bryan Taylor wrote:

> An incident is a form of ticket that recognizes that an existing requirement (customer or internal) isn't being met.
> 
> On 09/07/2011 06:20 AM, Soren Hansen wrote:
>> 2011/9/7 Bryan Taylor<btaylor at rackspace.com>:
>>> I'm working on an incident system that my company, as an OpenStack operator,
>>> will use to support our deployment of OpenStack.
>> 
>> Can you explain what you mean by "incidents"?
>> 
> 


Presumably you mean creating a support ticket when an error condition is reported by OpenStack? 
For Nova (compute) we have a notification api that reports various conditions. 

Nova itself would not interface with a ticketing (aka incident management) system.  The api's are all different, a fair amount of custom business logic would be needed to figure out what events in nova would constitute something 'incident' worthy, and it's really outside the domain of the compute service.  
Instead, you would have  a separate app that would subscribe to the notifications (they are published in Atom format and can be pushed via pubsubhubub.), make the business case decisions, and call whatever api the incident system has.  


--
	Monsyne M. Dragon
	OpenStack/Nova 
	cell 210-441-0965
	work x 5014190

This email may include confidential information. If you received it in error, please delete it.





More information about the Openstack mailing list