Socket timeouts while reading the producer response could indicate a
bottleneck on the server request handling. It could be in the network
layer, or i/o performance or config issue. It will help if you create a
JIRA and attach some part of your producer log that includes the timeout
errors. Also will be helpful to see the request log and server log of the
Kafka brokers for the time period of the timeouts.


On Wednesday, March 20, 2013, Bob Jervis 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