[
https://issues.apache.org/jira/browse/FLINK-8944?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16473346#comment-16473346
]
ASF GitHub Bot commented on FLINK-8944:
---------------------------------------
Github user tzulitai commented on a diff in the pull request:
https://github.com/apache/flink/pull/5992#discussion_r187788213
--- Diff:
flink-connectors/flink-connector-kinesis/src/main/java/org/apache/flink/streaming/connectors/kinesis/config/ConsumerConfigConstants.java
---
@@ -65,14 +65,14 @@ public SentinelSequenceNumber
toSentinelSequenceNumber() {
/** The date format of initial timestamp to start reading Kinesis
stream from (when AT_TIMESTAMP is set for STREAM_INITIAL_POSITION). */
public static final String STREAM_TIMESTAMP_DATE_FORMAT =
"flink.stream.initpos.timestamp.format";
- /** The base backoff time between each describeStream attempt. */
- public static final String STREAM_DESCRIBE_BACKOFF_BASE =
"flink.stream.describe.backoff.base";
--- End diff --
Changing the name of this variable, strictly speaking, breaks backwards
compatibility, as users might be using them.
> Use ListShards for shard discovery in the flink kinesis connector
> -----------------------------------------------------------------
>
> Key: FLINK-8944
> URL: https://issues.apache.org/jira/browse/FLINK-8944
> Project: Flink
> Issue Type: Improvement
> Reporter: Kailash Hassan Dayanand
> Priority: Minor
>
> Currently the DescribeStream AWS API used to get list of shards is has a
> restricted rate limits on AWS. (5 requests per sec per account). This is
> problematic when running multiple flink jobs all on same account since each
> subtasks calls the Describe Stream. Changing this to ListShards will provide
> more flexibility on rate limits as ListShards has a 100 requests per second
> per data stream limits.
> More details on the mailing list. https://goo.gl/mRXjKh
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
