Hi
peer-switch it will cause a recalculation at layer 2 as it makes both switches act as the root for STP , wouldn't be downtime per say but there may be a slight hiccup in paths at layer 2 until it settles
These commands are not really to be messing with during production time you should have in change request or maintenance window either way as something could go wrong during the calculation especially as its DC kit
If your using peer-switch with any type oif FHRP you should add peer-gateway too as best practice to allow load balancing across both 7ks at once
auto-recovery should not cause an issue , have not used arp sync but from what I understand it does I cant see it causing an issue the one I would be concerned about is peer-switch and setting that in a live network , it may be an instantaneous calculation but I didn't debug it when we applied it so I cant say exactly what went on in background
