Hi What are the best practises to achieve spooling support on producer end if the kafka cluster is not reachable or degraded.
We are thinking to have wrapper on kafka producer which can spool locally if the producer cant talk to kafka cluster. Problem with this approach is, all web servers which process requests will start having disks locally for spooling. Or is there any better way of doing this. I was going through this jira, looks like it is still open https://issues.apache.org/jira/browse/KAFKA-1955 - SunilKalva
