Feature Request: ability to keep node cordoned until specified

[imstillalive66]

Would like to propose an optional setting to keep nodes cordoned after restart and watch for an "uncordon-now" file to be written to uncordon the nodes.

Use case for instances where OS patching requires multiple restarts, in our case 2-3 times with varying uptimes between restarts.
This would allow tighter control over the nodes when we are within a limited patching window where setting a time to uncordon the node leaves them open to potential uncordoning of the node too soon allowing the next node to restart while the first node is still undergoing patching.

Happy to discuss options or alternatives if one exists!
Thanks!

[evrardjp]

This can be easier implemented in v2.

Side note: BTW you can already do that by manually cordon the node before kured events happen.
It will store cordoned state, so it will never uncordon until you say so.

Which means you can have your own controller that cordons/uncordons when you want, it will be simple, and will not mess with kured behaviour.