The infrastructure configuration uses Kubernetes through K3s across a three-node cluster. Containerized services give technical teams a repeatable packaging and deployment model, while Kubernetes manages placement and restart behavior.
Longhorn replicates persistent volume data. HAProxy and Keepalived provide a shared traffic entry point with failover configuration, and Traefik routes requests to the relevant application services. These mechanisms support resilience at different infrastructure layers.
Recovery depends on the service, available capacity and failure conditions. The current self-hosted MES database has one active instance per environment; its recovery can require a restart and volume reattachment.
Business valueLess dependence on individual components and a practical foundation for service recovery and managed maintenance.