Closed AlexandrChikur closed 1 year ago
I'm not sure what behaviour you're attempting to achieve exactly, could you try a more detailed example to help me out ?
Thanks, @dylanbstorey for reply! Im already dont need a help, thank you a lot.
I found some configs for confluent-kafka Producer class on which ProduceToTopicOperator based.
So I set kafka_config={"message.timeout.ms": "4000"}
and wrote my own delivery callback function, that raise AirflowException
if ack has an error.
I think thats a better solution honestly as it keeps the separation of concerns very clear.
Hi there!
I have some troubles with using the ProduceToTopicOperator in cases when Kafka is unavailable. The problem is the task is "hanging" for ... 30 or more minutes on producing. execution_timeout & poll_timeout does not anything
I'd like to limit task execution time, could anybody help me ?
Here is operator definition:
Producer function:
Task duration:
upd.: I did a research and now i know that it tooks 3 attempts with 5 mins on each and then mark task as failed. So i want to chage this behaviour to 4 sec on each attempt, and also mark it as failed task