From 9aa9f166b016217c4dc877600f5fa6b0e755dee0 Mon Sep 17 00:00:00 2001 From: Yash Sharma Date: Thu, 18 May 2017 22:38:43 +1000 Subject: [PATCH] Add documentation for kinesis retry configurations --- docs/streaming-kinesis-integration.md | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/docs/streaming-kinesis-integration.md b/docs/streaming-kinesis-integration.md index 6be0b548bc62b..9709bd3d6574d 100644 --- a/docs/streaming-kinesis-integration.md +++ b/docs/streaming-kinesis-integration.md @@ -216,3 +216,7 @@ de-aggregate records during consumption. - If no Kinesis checkpoint info exists when the input DStream starts, it will start either from the oldest record available (`InitialPositionInStream.TRIM_HORIZON`) or from the latest tip (`InitialPositionInStream.LATEST`). This is configurable. - `InitialPositionInStream.LATEST` could lead to missed records if data is added to the stream while no input DStreams are running (and no checkpoint info is being stored). - `InitialPositionInStream.TRIM_HORIZON` may lead to duplicate processing of records where the impact is dependent on checkpoint frequency and processing idempotency. + +#### Kinesis retry configuration + - `spark.streaming.kinesis.retry.waitTime` : Wait time between Kinesis retries as a duration string. When reading from Amazon Kinesis, users may hit `ProvisionedThroughputExceededException`'s, when consuming faster than 5 transactions/second or, exceeding the maximum read rate of 2 MB/second. This configuration can be tweaked to increase the sleep between fetches when a fetch fails to reduce these exceptions. Default is "100ms". + - `spark.streaming.kinesis.retry.maxAttempts` : Max number of retries for Kinesis fetches. This config can also be used to tackle the Kinesis `ProvisionedThroughputExceededException`'s in scenarios mentioned above. It can be increased to have more number of retries for Kinesis reads. Default is 3.