Design and defend a highly available ECS service by selecting EC2 or Fargate and specifying multi-AZ placement, load balancing, autoscaling, health checks, task IAM, and secrets; identify requirements that would justify EKS instead.

Diagnose a failed or unstable EKS request path using workload status, events, probes, Services, Ingress, autoscaling, disruption controls, and node signals.

Design an RDS topology that meets availability and recovery targets using Multi-AZ deployment, read replicas, backups, failover, connection management, and encryption.