Description
Problem Statement
I've been looking into the excellent foundation laid in #17617 and I think there's a great opportunity to extend this work to provide even more comprehensive monitoring capabilities, by exposing this on a non SQL endpoint like the jmx-exporter endpoint as a prometheus metric.
The current JMX exporter implementation provides valuable insights into CrateDB's performance, but there are several additional metrics that would be incredibly useful for production monitoring and troubleshooting. Building on the solid groundwork from #17617, we could expose more detailed metrics that would help operators better understand their cluster's behavior.
Possible Solutions
Following the pattern established in #17617, these could be implemented by:
Extending existing MBeans:
- Add new attributes to current MBean interfaces
- Creating new MBean categories: For metrics that don't fit existing structures
- I assume backporting this to 5.10 or 5.9 is off the table.
Considered Alternatives
No response