Improve debugging for brokered resources

[adracus]

Summary

machinebroker and volumebroker generate names for the brokered resources that might appear cryptic or lengthy in the
target cluster. There should be an easy way to track brokered objects across clusters.

Basic example

If a Machine gets created via machinebroker, the resulting entity might look like this:

my-namespace 12019901783616d84f7f1a806e948abc90a017c54eb3ac2450a0835a0322a23 my-class my-pool Running 1h

When created via machinepoollet, the broker annotations of the brokered machine contain the name, namespace and uid of the source machine.

Motivation

It is cumbersome to dive through the annotations / traverse multiple clusters to see how a machine got from a source to a target cluster. To debug easily and properly, there should be tooling to deal with the 'brokerage'.

For original issue, see #598