Metrics issue in zeebe cluster environment

Not precisely accurate in two senses:

First, there can only be one leader for a partition at any one time (unless you are seeing a bug), but one broker can be the leader for n partitions. So there is a 1:n relationship between broken:partition leader.

Secondly, this can change over time, so more than one broker can show up as the partition leader historically if the partition leadership changes. So the 1:n broker:partition leader constraint is valid only at a single point in time.

If those stats are over time, then they could be an accurate representation of the system.