[
https://issues.apache.org/jira/browse/FLINK-11617?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16768330#comment-16768330
]
Jamie Grier commented on FLINK-11617:
-------------------------------------
Here's an example:
Stacktrace is:
{{java.lang.RuntimeException: Rate Exceeded for getRecords operation - all 3
retry attempts returned ProvisionedThroughputExceededException.}}
{{ at
org.apache.flink.streaming.connectors.kinesis.proxy.KinesisProxy.getRecords(KinesisProxy.java:234)}}
{{ at
org.apache.flink.streaming.connectors.kinesis.internals.ShardConsumer.getRecords(ShardConsumer.java:373)}}
{{ at
org.apache.flink.streaming.connectors.kinesis.internals.ShardConsumer.run(ShardConsumer.java:216)}}
{{ at java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:511)
at java.util.concurrent.FutureTask.run(FutureTask.java:266)}}
{{ at
java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)
at
java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)}}
{{ at java.lang.Thread.run(Thread.java:748)}}
But the root cause is actually given by this log line:
{{Got recoverable SdkClientException. Backing off for 140 millis (null
(Service: AmazonKinesis; Status Code: 500; Error Code: InternalFailure; Request
ID: c49c8e5b-a068-9733-9043-b215d51b0aa1))}}
> Kinesis Connector getRecords() failure logging is misleading
> ------------------------------------------------------------
>
> Key: FLINK-11617
> URL: https://issues.apache.org/jira/browse/FLINK-11617
> Project: Flink
> Issue Type: Improvement
> Components: Kinesis Connector
> Affects Versions: 1.5.6, 1.6.3, 1.7.1
> Reporter: Jamie Grier
> Assignee: Jamie Grier
> Priority: Major
> Labels: pull-request-available
> Time Spent: 10m
> Remaining Estimate: 0h
>
> There isn't enough information in the current logging to diagnose a
> getRecords() failure. Also there is a hardcoded string that states the
> failure cause was always ProvisionedThroughputExceededException which isn't
> true. There are many possible causes of failures. This is misleading.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
