Add a metric to show the number of and size of the queue of node operations

[rkruze]

Who is this for and what problem do they have today?

When auto rebalance on node add is enabled, there may be many partition movements that happen on the cluster. Having visibility into the number of queued-up and number of executed operations like this as a metric would be helpful to understand if the cluster is still rebalancing or has finished.

What are the success criteria?

Provide two metrics, one for the number of operations that have been performed and another for the size of the queue of operations yet to be executed.

Why is solving this problem impactful?

To provide visibility into the partition movement on a cluster.

Additional notes