Add some sort of URL (www.https://[host]:[port]/health ) in JMX exporter to return 200 which indicate that server is working fine. We had a use case where we just wanted to check the health without scrapping the beans. Idea was to check if JMX exporter service is working fine and raise alarm if JMX exporter's health is bad.
Changes can be done here:
https://github.com/prometheus/jmx_exporter/blo

Currently Surface.of[X] automatically consume ClassNotFound exception and returns Surface.of[java.lang.Object] so that it will not stop user programs.

But as we hit the class loader issue like sbt/sbt#5410 and #918 this behavior makes difficult to identify what happened behind the scenes, so we should be able to report ClassNotFoundException later.

Another approac

Currently, the JMX port is completely unsecured. We should document how to use SSL and then use a username/password to connect to JMX.

The prometheus plugin uses metric names like this:

• org_graylog2_inputs_gelf_tcp_GELFTCPInput_5cc1bd20057c5200116dfdef_open_connections
• org_graylog2_lookup_adapters_5cc1bc798e510600109570d9_refresh_count
• org_graylog2_lookup_caches_5ca37df31210da0011143090_hits_total

It is really painful to write alarms for metrics like this. It would much better if the metrics would have names like th

As more traces are added to the traces table, its height increases and the Camel route diagram shrinks until it is no longer visible.

The update of jmxeval from 2.0 to version 3.x seems to require a migration as the 1.2 schema is gone. Sadly i cannot find a documentation for that. What do i have to change?