[
https://issues.apache.org/jira/browse/FLINK-6042?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17265851#comment-17265851
]
Till Rohrmann commented on FLINK-6042:
--------------------------------------
Any signature change to {{AccessExecutionGraph}} needed to let
{{ArchivedExecutionGraph}} return more {{ErrorInfos}} would also affect the
{{ExecutionGraph}}. Hence, it might be easier to introduce a {{JobInformation}}
object which is sent to the REST handlers. This object could contain the
{{AccessExecutionGraph}} and additional information (e.g. the exception
history).
Collecting and grouping the exceptions might be not super trivial. Keep in mind
that not every recovery is a global recovery and there can potentially be
concurrent recovery actions happening if we are running an embarrassingly
parallel job. Conceptually, I guess that one should start collecting the
exceptions whenever a failure occurs. This will then either trigger a recovery
or failing the job. If it is a recovery, then one would collect all exceptions
for the set of tasks to be restarted until one has restarted this set. If the
failure leads to a job failure, then we cancel the job and collect all
exceptions until the job has been canceled.
Since there can be concurrent recoveries happening, I am not sure whether the
{{JobStatusListener}} is sufficient because you wouldn't notice the second
recovery operation if you are already in the {{JobStatus.RESTARTING}}.
> Display last n exceptions/causes for job restarts in Web UI
> -----------------------------------------------------------
>
> Key: FLINK-6042
> URL: https://issues.apache.org/jira/browse/FLINK-6042
> Project: Flink
> Issue Type: Improvement
> Components: Runtime / Coordination, Runtime / Web Frontend
> Affects Versions: 1.3.0
> Reporter: Till Rohrmann
> Assignee: Matthias
> Priority: Major
>
> Users requested that it would be nice to see the last {{n}} exceptions
> causing a job restart in the Web UI. This will help to more easily debug and
> operate a job.
> We could store the root causes for failures similar to how prior executions
> are stored in the {{ExecutionVertex}} using the {{EvictingBoundedList}} and
> then serve this information via the Web UI.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
