I've read through many sites regarding using Kafka for log aggregation but haven't really found anything that actually talks about how people are shipping their logs into Kafka and consuming them. I'm really interested in implementation that would watch any kind of logs (local syslogs and application logs) and ship them near realtime into kafka. I think products like logstash and flume really shine in this area as they have multitude of options to ship any data stream into central aggregation service.

Since Kafka is proclaimed to be far more scalable I'm hoping there are options such as (http://logstash.net/docs/1.1.13/) to be able to vacuum any data source and put it into kafka queues and then consume them.

Any suggestions?