Github user pwendell commented on a diff in the pull request:
https://github.com/apache/spark/pull/5843#discussion_r30107017
--- Diff: core/src/main/scala/org/apache/spark/ui/jobs/StagePage.scala ---
@@ -434,25 +470,208 @@ private[ui] class StagePage(parent: StagesTab)
extends WebUIPage("stage") {
val content =
summary ++
showAdditionalMetrics ++
+ makeTimeline(stageData.taskData.values.toSeq, currentTime) ++
<h4>Summary Metrics for {numCompleted} Completed Tasks</h4> ++
<div>{summaryTable.getOrElse("No tasks have reported metrics
yet.")}</div> ++
<h4>Aggregated Metrics by Executor</h4> ++ executorTable.toNodeSeq
++
maybeAccumulableTable ++
<h4>Tasks</h4> ++ taskTable
-
UIUtils.headerSparkPage("Details for Stage %d".format(stageId),
content, parent)
}
}
+ def makeTimeline(tasks: Seq[TaskUIData], currentTime: Long): Seq[Node] =
{
--- End diff --
Can we make it so that we have some task count `MAX_TIMELINE_TASKS` and if
there are more tasks than that we don't generate and send this data? The issue
is that currently this UI has scalability issues because the scala XML output
can create very large amounts of intermediate data and actually cause the
driver to OOM. For now, maybe if there are more than 1000 or so tasks we should
not include this. Over time we can increase the count if we find that it is
okay.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
