About the EC2 persistent instance retirement scheduled

It's difficult to answer some of your questions because I don't know the reason for the instance retirement - for example, it could be that the hardware is impaired; or we could be removing older hardware to make way for newer hardware. You also haven't specified the instance type so that limits my answers to being general:

1. Instance storage is ephemeral and it is a customer responsibility to ensure that data on instance storage is backed up. As an example: If the hardware that your instance is running on were to suddenly fail and it was configured for instance recovery then the instance would be restarted on new hardware but the instance storage would need to be reinitialised by you. So yes, make sure your data is backed up and as per an answer I provided the other day you should check to see that the instance storage is initialised correctly at every operating system boot.
2. I can't answer that question - but the support team may be able to.
3. Without knowing the reason for the retirement I can't say. But if this event is about hardware impairment or retirement then creating a new instance should be ok. It's only if we're retiring an instance class then you'll need to choose a new instance type. I'd note that moving to newer instance types is a good idea anyway - you generally get higher performance for lower cost. You can move an Elastic IP from one instance to another; but not if it is a Public IP.

I'd also add that building resilience into your application on AWS (e.g. instance auto-recovery; auto-scaling; etc.) is a good thing (if possible). That way, events such as hardware failure, power outages, network isolation means that your environment can recovery automatically.