Explore how infrastructure resilience should follow workload criticality, redundancy, recovery, observability, and tested dependencies rather than a uniform high-availability label.