Are there really a few *hundred* metrics in Kafka 0.8 vs. a few dozen(?) in Kafka 0.7?
If so, I'll use something like https://github.com/sematext/jmxc
to get a JMX dump to see what's in each JMX.... but a few hundred metrics are going to be hard to map to the bullet points under http://kafka.apache.org/08/ops.html#monitoring
Performance Monitoring for Solr / ElasticSearch / HBase / Hadoop - http://sematext.com/spm