Placeholder issue for working on improving the logging behavior of pipeline deployments.
[ ] If the Python runner sidecar is not available, the pipeline should log an error and throw an exception. It should terminate with an exit code > 0.
[ ] If one of the records could not be successfully processed by the Python runner, the pipeline should log an error and throw an exception. It should NOT send the record to the stream out. If a dead letter queue is defined, the record should be sent to the dead letter queue. If no dead letter queue is defined, the processing should stop, allowing the developer to skip the record by resetting the offset of the pipeline or to continue the processing by adjusting the pipeline in a way such that it can process the record without throwing errors. At the moment, we do not support a dead letter queue, thus I suggest to start with the latter part.
Placeholder issue for working on improving the logging behavior of pipeline deployments.