Kafka: one producer for two topics vs. two producers - scala

Question

There are two kafka-topics:
logs (in text format), for this case I use standard StringSerializer for Kafka
events (in JSON format), for this case I use custom JSON Serializer for Kafka
There is some REST web-application (based on Servlet).
Which approach is best for this application?
Approach 1: Create single producer for both topics:
val producer = new KafkaProducer[String, AnyRef](...props...)
// send logs
producer.send(new ProducerRecord[String, AnyRef](
topic = "logs", "some log key", "some log str"))
// send events
producer.send(new ProducerRecord[String, AnyRef](
topic = "events", "some evt key", Event("some"))
Approach 2: Create two producers with strong types of values.
val logsProducer = new KafkaProducer[String, String](...props...)
val eventsProducer = new KafkaProducer[String, Event](...props...)
// send logs
logsProducer.send(new ProducerRecord[String, String](
topic = "logs", "some log key", "some log str"))
// send events
eventsProducer.send(new ProducerRecord[String, Event](
topic = "events", "some evt key", Event("some event"))
Update 1: For the Approach-1 I use own serializer based on Json4s:
class KafkaJson4sSerializer[T <: AnyRef] extends Serializer[T] {
import org.json4s._
import org.json4s.native.Serialization
import org.json4s.native.Serialization.write
implicit val formats = Serialization.formats(NoTypeHints)
override def configure(configs: util.Map[String, _], isKey: Boolean): Unit = {}
override def serialize(topic: String, data: T): Array[Byte] = {
write(data).getBytes
}
override def close(): Unit = {}
}
val p = new Properties()
p.put(ProducerConfig.KEY_SERIALIZER_CLASS_CONFIG,
"org.apache.kafka.common.serialization.StringSerializer")
p.put(ProducerConfig.VALUE_SERIALIZER_CLASS_CONFIG,
"KafkaJson4sSerializer") // use the above own serializer
val producer = new KafkaProducer[String, AnyRef](p)
// send string type to topic 'logs'
producer.send(new ProducerRecord[String, AnyRef]("logs", "k1", "string value"))
// send Event type to another topic 'events'
producer.send(new ProducerRecord[String, AnyRef]("events", "k2", Event("some evt")))