edwinalu commented on a change in pull request #23767: [SPARK-26329][CORE][WIP]
Faster polling of executor memory metrics.
URL: https://github.com/apache/spark/pull/23767#discussion_r257375347
##########
File path: core/src/main/scala/org/apache/spark/executor/Executor.scala
##########
@@ -189,9 +203,44 @@ private[spark] class Executor(
*/
private val HEARTBEAT_INTERVAL_MS = conf.get(EXECUTOR_HEARTBEAT_INTERVAL)
+ /**
+ * Interval to poll for executor metrics, in milliseconds
+ */
+ private val METRICS_POLLING_INTERVAL_MS =
conf.get(EXECUTOR_METRICS_POLLING_INTERVAL)
+
+ // Executor for the metrics polling task
+ private val poller =
+ if (METRICS_POLLING_INTERVAL_MS > 0) {
+ ThreadUtils.newDaemonSingleThreadScheduledExecutor("executor-poller")
+ } else {
+ null
+ }
+
+ private def poll(): Unit = {
Review comment:
Having the logic for tracking the peak values for executor metrics in
Executor.scala is exposing a lot of the implementation details for
ExecutorMetrics (Array[Long]). Admittedly Array[Long] is already used for
tests, but it would be nice to maintain more of the abstraction. Perhaps a
ConcurrentExecutorMetrics?
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
With regards,
Apache Git Services
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
