If configured properly, Kafka should garbage collect objects fairly
regularly since request/response objects are typically short lived. It
seems that the only thing that would cause memory usage increase
proportional to the number of topics is metrics. One issue with garbage
collection in Kafka is the request purgatory, that is if the consumer
clients use very large request timeouts, the fetch request objects would
sit in the internal purgatory data structure for long time without being
garbage collected. Do you have consumer clients that have set a large value
for "" ?

On Mon, Sep 2, 2013 at 10:54 PM, Vadim Keylis <[EMAIL PROTECTED]> wrote:
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