When an Amazon Neptune workload needs resilience beyond a single cluster, this post presents a decision framework that maps availability targets to Neptune architectural patterns, from Multi-AZ and Global Database to write-ahead logs and caching, based on your RTO, RPO, and budget.