The image shows an architecture diagram of a multiregion Kubernetes disaster recovery configuration.

There are two regions: Region 1 - Primary and Region 2 - Secondary. A Domain Name Service (DNS), a kubernetes application (myk8sapp.example.com), and a Container Registry for other external container registries are outside of the regions. There is two-way communication between the DNS and myk8sapp.example.com.

Each region has Load Balancer, a Kubernetes Cluster, a Container Registry, an Oracle RAC Database, ETCD Backups, YAML Snapshots, a Kube API, and a dynamic routing gateway (DRG). There is two-way communication between the DRGs in each region. The Load Balancer in Region 1 and Region 2 are connected to myk8sapp.example.com.

The Kubernetes Cluster in each region contains the following:

Ingress Controller: Data flows from the Ingress Controller to the Load Balancer.
Kubernetes Worker Nodes with 3 nodes (ci-w1, ci-w2, and ci-w3). The Kubernetes Worker Nodes are connected to the Ingress Controller and to Oracle RAC Database.
Kubernetes Control Plane with three control planes (ci-cp1, cci-cp2, and ci-cp3) and three ETCD instances. The Kubernetes Control Plane nodes are connected to the Kube API.

In Region 1, data from Oracle RAC Database flows through Oracle Data Guard to the Oracle RAC Database in Region 2.

In Region 1, the Control Plane uses the Kube API to send snapshots to YAML Snapshots. Region 1 YAML Snapshots flow to Region 2 YAML Snapshots and then to Region 2's Kube API.

In Region 1, the Control Plane sends backups to ETCD Backups. Region 1 ETCD Backups flow to Region 2 ETCD Backups and then to Region 2's Kubernetes Control Plane.