You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@spark.apache.org by "Apache Spark (JIRA)" <ji...@apache.org> on 2018/05/28 03:35:00 UTC
[jira] [Assigned] (SPARK-24398)
SQLAppStatusListener.aggregateMetrics() invocation is too slow
[ https://issues.apache.org/jira/browse/SPARK-24398?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Apache Spark reassigned SPARK-24398:
------------------------------------
Assignee: (was: Apache Spark)
> SQLAppStatusListener.aggregateMetrics() invocation is too slow
> ---------------------------------------------------------------
>
> Key: SPARK-24398
> URL: https://issues.apache.org/jira/browse/SPARK-24398
> Project: Spark
> Issue Type: Improvement
> Components: Web UI
> Affects Versions: 2.3.0
> Reporter: furongbin
> Priority: Major
>
> RROR 18/05/24 14:08:25 AsyncEventQueue: Dropping event from queue appStatus. This likely means one of the listeners is too slow and cannot keep up with the rate at which tasks are being started by the scheduler. [dag-scheduler-event-loop]
> WARN 18/05/24 16:39:54 AsyncEventQueue: Dropped 208205 events from appStatus since Thu May 24 16:30:57 CST 2018. [dag-scheduler-event-loop]
> WARN 18/05/24 16:40:19 SQLAppStatusListener: agregateMetrics cost=18.6s, metricIds.size=775, metrics.size=5292940, aggregatedMetrics=615 [pool-22-thread-1]
> private def aggregateMetrics(exec: LiveExecutionData): Map[Long, String] = {
> val metricIds = exec.metrics.map(_.accumulatorId).sorted
> val metricTypes = exec.metrics.map
> { m => (m.accumulatorId, m.metricType) }
> .toMap
> val metrics = exec.stages.toSeq
> .flatMap \{ stageId => Option(stageMetrics.get(stageId)) }
> .flatMap(_.taskMetrics.values().asScala)
> .flatMap \{ metrics => metrics.ids.zip(metrics.values) }
> {color:#ff0000}val aggregatedMetrics = (metrics ++ exec.driverAccumUpdates.toSeq){color}
> {color:#ff0000} .filter \{ case (id, _) => metricIds.contains(id) }{color}
> .groupBy(_._1)
> .map
> { case (id, values) => id -> SQLMetrics.stringValue(metricTypes(id), values.map(_._2).toSeq) }
> if (exec.metricsValues != null)
> { exec.metricsValues }
> else
> { aggregatedMetrics }
> }
> *Case:*
> when aggregatedMetrics.size=5000000, metrics.size=1000, then the execution({color:#ff0000}val aggregatedMetrics = (metrics ++ exec.driverAccumUpdates.toSeq){color}
> {color:#ff0000}.filter \{ case (id, _) => metricIds.contains(id) }{color}) is very slow
>
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@spark.apache.org
For additional commands, e-mail: issues-help@spark.apache.org