Zero-Downtime AWS Kubernetes Architecture

I need a well-thought-out Kubernetes architecture on AWS that lets my microservices run without a single second of downtime. I’m already committed to EKS, and the cluster must integrate cleanly with RDS; everything from connection handling to automated credential rotation should be covered.

High availability is non-negotiable. Please bake in automatic failover, multi-AZ worker nodes, and a cross-region disaster-recovery plan so that traffic keeps flowing even during a full-region outage. Wherever possible, I’d like to leverage managed offerings—think Amazon EKS blue/green deployments, load-balancer health checks, and Route 53 weighted routing—to keep rollouts smooth.

Deliverables
• An architecture diagram (Visio, Draw.io, or Lucidchart) that shows VPC layout, subnets, node groups, RDS placement, and failover paths
• A concise strategy document describing:
– Deployment pipeline (GitHub Actions or CodePipeline)
– Zero-downtime rollout method (e.g., canary or blue/green)
– Backup, restore, and cross-region replication steps
– Monitoring stack using CloudWatch metrics and alerts
• Terraform or CloudFormation snippets for the critical pieces (EKS, node groups, RDS, networking)
• A short walkthrough call or recorded demo so I can understand and reproduce the setup

I value clarity over quantity—show me exactly how your design keeps services alive through node, AZ, or region failures, and we’re in business.