An enterprise retail organization is planning a Google Kubernetes Engine (GKE) cluster architecture to host an event-driven inventory processing system. The architecture must satisfy the following requirements:
- Worker nodes must reside in private subnets with no public IP addresses assigned to them.
- Nodes must be able to reach external third-party vendor APIs over the internet using a fully managed Google Cloud service.
- Access to the GKE control plane master endpoint must be restricted exclusively to designated management IP addresses.
- Background batch processing workloads are stateless and fault-tolerant, and must run on compute capacity that minimizes infrastructure costs.
Which GKE cluster architecture and configuration should you select to fulfill these requirements?
- Provision a Private GKE cluster with Master Authorized Networks enabled, deploy Cloud NAT within the VPC network for outbound traffic, and create a dedicated node pool using Spot VMs for batch processing.Answer
- BProvision a Private GKE cluster with Master Authorized Networks enabled, rely on the Horizontal Pod Autoscaler (HPA) to provision new Compute Engine instances for outbound connectivity, and disable Cloud NAT.
- CProvision a Public GKE cluster with Master Authorized Networks disabled, deploy Cloud NAT, and create a Spot VM node pool dedicated to hosting stateful persistent database storage.
- DProvision a GKE Standard cluster in public mode, configure custom sysctl parameters on private nodes, and rely on Cluster Autoscaler to dynamically provision third-party NAT gateways for outbound traffic.