We should start looking at consolidating our aarch64 workload (CI and cbs/koji) to some nodes and so free some rack space in the current DC where our infra is mostly hosted.
That would mean also then be able to get some refreshed hardware

Tasks:

• aarch64 up2date inventory (machines and workload)
• design optimized power/cpu usage (and space usage)
• migrate workload/Virtual Machines
• decommission (very) old aarch64 machines