Change wrapper consumer loop to check whether a KafkaError is marked as fatal before giving up, instead of giving up on any otherwise unhandled error. This might solve #86 (but can't really test other than on a live stream), but is probably worth doing anyway.
Change wrapper consumer loop to check whether a KafkaError is marked as fatal before giving up, instead of giving up on any otherwise unhandled error. This might solve #86 (but can't really test other than on a live stream), but is probably worth doing anyway.