Github user JasonMWhite commented on a diff in the pull request:
https://github.com/apache/spark/pull/10089#discussion_r46449571
--- Diff:
external/kafka/src/main/scala/org/apache/spark/streaming/kafka/DirectKafkaInputDStream.scala
---
@@ -89,23 +89,29 @@ class DirectKafkaInputDStream[
private val maxRateLimitPerPartition: Int =
context.sparkContext.getConf.getInt(
"spark.streaming.kafka.maxRatePerPartition", 0)
- protected def maxMessagesPerPartition: Option[Long] = {
+ protected def maxMessagesPerPartition(leaderOffsets:
Map[TopicAndPartition, LeaderOffset])
+ : Option[Map[TopicAndPartition, Long]] = {
val estimatedRateLimit = rateController.map(_.getLatestRate().toInt)
- val numPartitions = currentOffsets.keys.size
-
- val effectiveRateLimitPerPartition = estimatedRateLimit
- .filter(_ > 0)
- .map { limit =>
- if (maxRateLimitPerPartition > 0) {
- Math.min(maxRateLimitPerPartition, (limit / numPartitions))
- } else {
- limit / numPartitions
+
+ // calculate a per-partition rate limit based on current lag
+ val effectiveRateLimitPerPartition = estimatedRateLimit.filter(_ > 0)
match {
+ case Some(rate) =>
+ val lagPerPartition = leaderOffsets.map { case (tp, lo) =>
+ tp -> Math.max(lo.offset - currentOffsets(tp), 0)
+ }
+ val totalLag = lagPerPartition.values.sum.toFloat
+
+ lagPerPartition.map { case (tp, lag) =>
+ tp -> Math.round(lag / Math.max(totalLag, 1) * rate)
--- End diff --
Good point!
---
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]
