Hey all -- we have a flow where we read from Kinesis, but sometimes the connection times out. From what I can tell, it causes the job to restart. We have 10 restarts configured with a 60 second delay, but does Flink provide any additional configuration to handle Kinesis errors? Is there any way to handle timeouts that doesn't involve auto-restarting the job?

org.apache.flink.kinesis.shaded.com.amazonaws.SdkClientException: Unable to execute HTTP request: Connect to kinesis.us-west-2.amazonaws.com:443 [kinesis.us-west-2.amazonaws.com/34.223.45.32] failed: connect timed out

Not yet, but we are working on adding something. Please add a comment to this ticket, especially if the requirements/proposal would not meet your requirements: https://issues.apache.org/jira/browse/FLINK-33260