Steps to reproduce:

1. Create discovery rule for memory used with key:

   jmx.discovery[beans,"org.apache.flink.jobmanager.Status.JVM.Memory.Direct.MemoryUsed:*"]
   

2. Create Item prototype with key:

   jmx[{#JMXOBJ}, "Value"]
   

3. Create second discovery rule for another jmx metrics (i.e. CPU load) with key:

   jmx.discovery[beans,"org.apache.flink.jobmanager.Status.JVM.CPU.Load:*"]
   

4. Try to create item prototype for second discovery rule with key:

   jmx[{#JMXOBJ}, "Value"]
   

Result:

Error: "Element with the same key already exist..."

Expected:

No error.

If I add space for second item prototype key it works, but i have about 20 metrics there, so the last one will contain 20 spaces and that is ugly.

Is there another workaround to solve this?

P.S. I need multiple discovery rules because i'll need different trigger prototypes for theese metrics.