Not really - well, brokers obviously have to maintain these
connections from producers but it maintains no additional per-producer
Same for consumers. For high-level consumers, the consumers themselves
maintain state (e.g., registration, partition ownership, offsets) in
zookeeper but there is no per-consumer state on the broker (although
with KAFKA-1000, we will be moving offset management to the broker).



NEW: Monitor These Apps!
elasticsearch, apache solr, apache hbase, hadoop, redis, casssandra, amazon cloudwatch, mysql, memcached, apache kafka, apache zookeeper, apache storm, ubuntu, centOS, red hat, debian, puppet labs, java, senseiDB