3.1.1. Designing for High Availability & Disaster Recovery

An entire AWS Availability Zone went dark during the 2017 S3 outage — and companies that had designed for single-AZ architectures went dark with it. Without high availability design, a single hardware failure, network partition, or data center issue can take your entire application offline. The difference between a 5-minute blip and a 5-hour outage comes down to decisions you make before the failure happens.

Think of HA design like building a bridge with redundant cables — any single cable can fail without the bridge collapsing. This section covers Multi-AZ and Multi-Region deployment strategies, SLA translation, replication patterns, and failover methods. How many simultaneous component failures can your architecture survive?