Prepare and test your skills

Prepare and test your skills

Free sample question with answers and detailed explanation below. The Professional Cloud Network Engineer practice bank has an additional 2,437 practice questions.Try 10 free realistic questions: instant score, no credit card, and no account required.
An enterprise is migrating a legacy internal API from an on-premises data center to a VPC-native GKE cluster in us-central1 while deploying a new Cloud Run microservice in the same region. Your regional internal Application Load Balancer must satisfy these requirements:
/legacy/* requests across GKE Pod IPs and on-premises endpoints (10.250.10.0/24) within a single backend service connected via Cloud Interconnect in us-central1./v2/* requests to the Cloud Run service.Which NEG types and VPC routing configurations should you implement?
This design integrates three specialized Network Endpoint Group types behind a regional internal Application Load Balancer: a standalone zonal NEG (GCE_VM_IP_PORT) for container-native GKE Pod targeting, a hybrid connectivity NEG (NON_GCP_PRIVATE_IP_PORT) for private on-premises endpoints, and a serverless NEG (SERVERLESS) for Cloud Run.
GCE_VM_IP_PORT) can be combined with hybrid connectivity NEGs (NON_GCP_PRIVATE_IP_PORT) inside a single backend service, enabling seamless traffic distribution and gradual migration for /legacy/* via gcloud or the REST API.SERVERLESS) in the same region attaches to a dedicated backend service mapped to /v2/* in the load balancer's URL map.130.211.0.0/22 and 35.191.0.0/16) for zonal NEGs, permitting both ranges on GKE node firewalls (alongside permitting the proxy-only subnet on-premises) ensures every endpoint passes health checks.kube-proxy node hop.GCE_VM_IP_PORT and NON_GCP_PRIVATE_IP_PORT NEGs use the RATE balancing mode on Application Load Balancers.Unlike Ingress-managed NEGs, GCE_VM_IP passthrough NEGs, or regional internet NEGs, this combination satisfies Google Cloud's strict mixed-backend rules, preserves BGP dynamic routing across Cloud Interconnect, and isolates the serverless backend cleanly through URL-based routing.
This configuration attempts to combine a GKE Ingress-managed zonal NEG (GCE_VM_IP_PORT) with a hybrid connectivity NEG (NON_GCP_PRIVATE_IP_PORT) while using regional dynamic routing and relying exclusively on distributed Envoy health checks from the proxy-only subnet.
130.211.0.0/22 and 35.191.0.0/16), so relying only on proxy-only subnet probes causes GKE Pod health checks to fail.gcloud or the API results in configuration drift or reversion by the GKE controller.This approach would be partially applicable if the GKE workload were exposed independently via a dedicated GKE Ingress without any on-premises hybrid NEG mixed into the same backend service, and if global dynamic routing were enabled.
Even when using a regional load balancer, hybrid connectivity NEGs mandate global dynamic routing on the Cloud Router and require advertising the proxy-only subnet so on-premises firewalls and routers can route Envoy return traffic.
This configuration tries to use a GCE_VM_IP zonal NEG for GKE nodes and a regional internet NEG (INTERNET_FQDN_PORT) paired with a private Cloud DNS zone to reach on-premises endpoints inside a single /legacy/* backend service.
GCE_VM_IP endpoints represent primary VM network interface IP addresses without ports and are supported only by passthrough Network Load Balancers. Application Load Balancers require GCE_VM_IP_PORT endpoints to perform Layer 7 proxying and container-native Pod load balancing.GCE_VM_IP_PORT zonal NEGs with NON_GCP_PRIVATE_IP_PORT hybrid connectivity NEGs.GCE_VM_IP_PORT were used, the configuration omits firewall rules allowing Google's centralized health check ranges (130.211.0.0/22 and 35.191.0.0/16) to reach the GKE nodes.GCE_VM_IP constraints: Cannot target Pod alias IP ranges or custom container ports, and cannot attach to an INTERNAL_MANAGED backend service.INTERNET_FQDN_PORT via private Cloud DNS is a valid technique for reaching private on-premises backends at a regional rather than zonal scope, but it must reside in its own dedicated backend service.Using a GCE_VM_IP NEG would be appropriate for an internal passthrough Network Load Balancer, and using a regional internet NEG (INTERNET_FQDN_PORT) with private Cloud DNS would work if on-premises endpoints required regional availability in a separate backend service without mixing zonal NEGs.
Remember that Application Load Balancers and proxy Network Load Balancers always require GCE_VM_IP_PORT for zonal NEGs, whereas GCE_VM_IP is reserved exclusively for passthrough Network Load Balancers.
This configuration selects the correct NEG types—a standalone zonal NEG (GCE_VM_IP_PORT), a hybrid connectivity NEG (NON_GCP_PRIVATE_IP_PORT), and a serverless NEG (SERVERLESS)—but attempts to use static VPC routes and centralized Google health check probes for the on-premises endpoints.
NON_GCP_PRIVATE_IP_PORT endpoints.10.250.10.0/24 originates from Envoy proxies inside the proxy-only subnet, not from 130.211.0.0/22 or 35.191.0.0/16.130.211.0.0/22 and 35.191.0.0/16) are used for hybrid NEGs only on GFE-based global and classic load balancers (global external ALB, classic ALB, global external proxy NLB, classic proxy NLB).gcloud or the REST API rather than the Google Cloud console.Permitting 130.211.0.0/22 and 35.191.0.0/16 on the on-premises firewall and advertising those centralized ranges via BGP would be required if you were deploying a global external Application Load Balancer instead of a regional internal Application Load Balancer—though global dynamic routing on Cloud Router would still be mandatory.
Distinguishing between GFE-based centralized health checks and Envoy-based distributed health checks is critical when designing firewall rules and BGP advertisements for hybrid NEGs.
Answer them yourself and get your score the moment you finish. No account needed.
Full access includes
Keep the momentum going with these hand-picked practice scenarios
Last updated Validated