Features

As Grafana now supports derivatives the dashboards have been updated accordingly. This means for example that in the Host dashboard you see the actual Disk I/O per second (you need Grafana 2.6.1 for that) and not the bytes read and written since the last reboot. Other updated graphs include Network I/O and Garbage Collection times.
The rate at which you want to report request traces to stagemonitor can be controlled via stagemonitor.requestmonitor.onlyReportNRequestsPerMinuteToElasticsearch. Set to 0 or below to deactivate elasticsearch request trace reporting and to Integer.MAX_VALUE to always report (default).
Request traces and metrics can now be logged in Elasticearch Bulk format. That way nothing is directly sent over the network. You can use logstash or a different external process to send the information to ES.
The call tree view in the in browser widget now shows the query count. This is extremely useful to find out under which branch of the call tree Elasticsearch requests or SQL queries are buried. Thx @gedankennebel

Bug Fixes

Fixes the JMX reporter and the RegexMetricFilter (big thanks to @ryanrupp)
Fixes a memory leak that occurs on redeploy (see #119 thx @ryanrupp for reporting)

The hot-cold Elasticsearch architecture now has to be explicitly activated via setting stagemonitor.elasticsearch.hotColdArchitecture.moveToColdNodesAfterDays. Previously, if you didn't tag your nodes with box_type: <hot|cold> in elasticsearch.yml, shards could not have been allocated.