createKafkaWrite will fail indefinitely if kafka is unreachable at producer creation

[atamon]

Typically cumbersome for e2e tests with build-contract. Your service starts before kafka is ready -> you create a producer which tries, and fails, to connect to kafka, and then you have a worthless write function

[solsson]

I wonder if this is the expected behavior with librdkafka/node-rdkafka? If so we'll need to catch an error and retry with backoff I guess.