How to recover Ec2 Instance in failure condition

With an RPO of 30 minutes, it would be difficult to use EC2 snapshots and backups alone.
For example, important data should not be stored in EBS, but managed by an external, redundant service such as S3 or EFS.
For example, in the following document, if you use warm standby to replicate a database, it would be easy to spread the data to 30 minutes ago.
https://aws.amazon.com/jp/blogs/architecture/disaster-recovery-dr-architecture-on-aws-part-i-strategies-for-recovery-in-the-cloud/

I'd encourage you to use autoscaling group behind ELB so that such events can be handled automatically at least for EC2 instance, if EC2 instance is the only ask.

If there is anything related to data, then RTO, RPO come into the play and you would need to create DR plan based on which approach would work best for you out of those four:

1. Back up and Restore
2. Pilot Light
3. Warm Standby
4. Active-Active

These are ordered in ascending way for cost and descending way for RTO/RPO.

RTO/RPO: 4<3<2<1 (1 would have longest outage duration) Cost: 4>3>2>1(Maximum cost to achieve 4)

Alongwith blog post shared earlier in one of the answer, here are some of the re:Invent videos, which are really helpful:

Architecture Patterns for Multi-Region Active-Active Applications Designing for failure: Architecting resilient systems on AWS