[GitHub] [spark] otterc commented on a diff in pull request #36165: [SPARK-36620][SHUFFLE] Add Push Based Shuffle client side metrics

[GitBox <gi...@apache.org>]

otterc commented on code in PR #36165:
URL: https://github.com/apache/spark/pull/36165#discussion_r1030746682


##########
core/src/main/scala/org/apache/spark/storage/ShuffleBlockFetcherIterator.scala:
##########
@@ -282,6 +280,17 @@ final class ShuffleBlockFetcherIterator(
       }
     }
 
+    @inline def updateMergedReqsDuration(wasReqForMergedChunks: Boolean = false): Unit = {
+      if (remainingBlocks.isEmpty) {
+        val durationMs = TimeUnit.NANOSECONDS.toMillis(clock.nanoTime() - requestStartTime)
+        if (wasReqForMergedChunks) {
+          shuffleMetrics.incRemoteMergedReqsDuration(durationMs)
+        } else {
+          shuffleMetrics.incRemoteReqsDuration(durationMs)
+        }

Review Comment:
   We added these metrics internally so we can easily figure out what is taking longer- the fetch of the merge chunks or the few unmerged blocks. If we include the duration of merged chunks in the total then we have to derive the value for un-merged blocks which is fine. It makes sense to be consistent to avoid confusion.



-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org