[
https://issues.apache.org/jira/browse/FALCON-1205?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14564281#comment-14564281
]
Pallavi Rao commented on FALCON-1205:
-------------------------------------
I see. Thanks [~ajayyadava] for clarifying. The state you are persisting will
suffice, if you are checking for pending instances in a window and publishing
any SLA misses in that window, as a batch, to a JMS endpoint. But, if you need
to satisfy queries like this - "Give me all feed instances that missed their
SLA in the last 2 hours (random time duration), via a REST endpoint. The stored
state in its current form will not help. You'll end up scanning the instances
for the entire range specified by the user, anyway, right? May be, if we change
the schema of the stored state and store it also as periodic datasets, we won't
need to recompute for REST queries?
Sorry for being naive. If you can put the overall design in a few words, it
will help me understand.
> SLAService to keep track of missing SLAs for feeds
> --------------------------------------------------
>
> Key: FALCON-1205
> URL: https://issues.apache.org/jira/browse/FALCON-1205
> Project: Falcon
> Issue Type: Sub-task
> Reporter: Ajay Yadava
> Assignee: Ajay Yadava
> Attachments: FALCON-1205.patch
>
>
> We need a service to keep track of missing feed instances.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
