Features

Stagemonitor now uses P6Spy3 for JDBC monitoring and SQL recording which offers a new event based architecture which has less overhead and allows for a simpler implementation inside stagemonitor.
- This also allows for data source monitoring without bytecode instrumentation. Because of this, data sources like druid which can't be monitored with bytecode manipulation can now be monitored. See #201
Stagemonitor now doesn't convert durations on histograms as the histogram might not contain time information but rather things like response size bytes. Thx @captSL
Fix Kibana visualizations/dashboards to work with version 5. Thx @domtoupin
Adds support to "profile" Freemarker template files. This feature answers the question "which FTL file (and line number) caused the call of a expensive method"? See #195
Logging improvements. See #198. Thx @domtoupin
Stagemonitor now periodically checks if Elasticsearch is available. If not, it does not try to send request traces to ES which would fill up the request queue and then eventually fail. See #208
- Set default HTTP connection timeout to 5s and read timeout to 15s
Added the ability to set number of shards and replicas of ES indices

Bug fixes