[
https://issues.apache.org/jira/browse/FALCON-1495?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
sandeep samudrala updated FALCON-1495:
--------------------------------------
Description:
Most of the times , when there are failures in the pipelines, retries run and
the workflows might have got succeeded in sub sequential retry. But for end
user to figure out why it has failed initially might be useful to understand
for failures handling and SLA misses.
As of now to figure out the same, its a tedious task to get such information by
either going through oozie DB/logs or via the logs backed up in hdfs by the
falcon postprocessing(which is 100% reliable as in few cases, falcon post
processing might fail or for some reason if it couldn't back up the logs).
In either case, Falcon should abstract the oozie client layer to be used by end
user, and support all runs status to end user.
was:
Most of the times , when there are failures in the pipelines, retries run and
the workflows might have got succeeded in sub sequential retry. But for end
user to figure out why it has failed initially/ or in case why there was a SLA
breach.
As of now to figure out the same, its a tedious task to get such information by
either going through oozie DB/logs or via the logs backed up in hdfs by the
falcon postprocessing(which is 100% reliable as in few cases, falcon post
processing might fail or for some reason if it couldn't back up the logs).
In either case, Falcon should abstract the oozie client layer to be used by end
user, and support all runs status to end user.
This task will have to add an extra optional Boolean flag to instance list
(-allRuns), which will return all runs for given instances, from which user can
further look for information related to which sub action in user workflow has
failed and for what reason.
> In instance status list, show all runs for instances when requested by user
> ---------------------------------------------------------------------------
>
> Key: FALCON-1495
> URL: https://issues.apache.org/jira/browse/FALCON-1495
> Project: Falcon
> Issue Type: Sub-task
> Components: ease, general
> Reporter: sandeep samudrala
>
> Most of the times , when there are failures in the pipelines, retries run and
> the workflows might have got succeeded in sub sequential retry. But for end
> user to figure out why it has failed initially might be useful to understand
> for failures handling and SLA misses.
> As of now to figure out the same, its a tedious task to get such information
> by either going through oozie DB/logs or via the logs backed up in hdfs by
> the falcon postprocessing(which is 100% reliable as in few cases, falcon post
> processing might fail or for some reason if it couldn't back up the logs).
> In either case, Falcon should abstract the oozie client layer to be used by
> end user, and support all runs status to end user.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
