org.apache.spark.streaming.kafka.KafkaCluster
Make a consumer config without requiring group.id or zookeeper.connect, since communicating with brokers also needs common settings such as timeout