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.
Your organization is deploying self-managed multi-NIC network virtual appliances (NVAs) across multiple zones in us-west1 to inspect internal VPC traffic. Each NVA has nic0 attached to untrust-vpc and nic1 attached to trust-vpc. You are configuring an internal passthrough Network Load Balancer in trust-vpc as the next hop of a custom static route to distribute traffic to nic1 with the following requirements:
nic1 interface so that an interface or VM failure automatically redistributes traffic to healthy appliances.Which configuration meets all of these requirements?
GCE_VM_IP zonal NEGs across multiple zones targeting nic1 on an internal backend service with CLIENT_IP session affinity, failover backends with connection draining, and a regional TCP-based health check with logging enabled. Allow ingress from 35.191.0.0/16 to nic1, bind the NVA health-check listener to 0.0.0.0, and configure policy-based routing (ip rule add from <LB_VIP>/32) so responses exit via nic1.This architecture pairs an internal passthrough Network Load Balancer with GCE_VM_IP zonal network endpoint groups (NEGs), policy-based routing, TCP-based health checks, and failover controls to inspect traffic across multi-NIC virtual appliances while maintaining high availability and session continuity.
nic1 targeting and health monitoring: Unlike instance groups—which always send traffic and probes to nic0—GCE_VM_IP zonal NEGs bind directly to the subnet of nic1 in trust-vpc. Allowing ingress from 35.191.0.0/16 permits Google Cloud health check probers to reach nic1. Because passthrough load balancers preserve the load balancer virtual IP (VIP) as the destination address, binding the listener to 0.0.0.0 accepts the probes, and the Linux policy routing rule (ip rule add from <LB_VIP>/32) forces reply packets to egress via nic1 rather than the default nic0 route.gcloud compute backend-services get-health and the Google Cloud console expose real-time HEALTHY and UNHEALTHY interface states.CLIENT_IP session affinity keeps inspection flows pinned to the same NVA during normal operation. Configuring failover backends and connection draining on failover ensures that if nic1 or an entire NVA fails consecutive health checks, traffic automatically redistributes to healthy appliances with minimal disruption.nic0 support: GCE_VM_IP zonal NEGs natively target secondary interfaces across multiple zones.nic0 and nic1.Instance groups cannot target nic1, proxy load balancers cannot act as custom static route next hops, and omitting policy-based routing causes health check probes to fail silently due to return traffic exiting nic0.
CLIENT_IP session affinity and a failover policy. Attach a regional HTTP health check with logging enabled, allow ingress from 35.191.0.0/16 in trust-vpc, and configure the NVA health-check daemon to listen exclusively on nic1's primary internal IP address.This configuration attempts to use a regional managed instance group (MIG) with autoscaling and failover settings behind an internal passthrough Network Load Balancer, while restricting the NVA health-check service to listen only on nic1's assigned primary internal IP address.
nic1: In Google Cloud, instance group backends (both managed and unmanaged) only receive load-balanced traffic and health check probes on the VM's primary interface (nic0). Because nic0 resides in untrust-vpc, an instance group cannot distribute traffic or monitor nic1 in trust-vpc.nic1's primary internal IP. A daemon bound exclusively to nic1's IP drops these probes.nic0.0.0.0.0 or the VIP.ip rule matching the VIP as source IP, outbound health check replies still exit via nic0.The load balancer targeted nic0 instead of nic1, autoscaling were disabled when using failover backends, policy-based routing were not needed, and the NVA listener bound to 0.0.0.0 or the load balancer VIP.
To load balance or health-check any non-nic0 interface on a multi-NIC VM, you must replace instance groups with GCE_VM_IP zonal NEGs in the target subnet.
GCE_VM_IP_PORT zonal NEGs targeting nic1 in trust-vpc, CLIENT_IP session affinity, and a regional TCP health check with logging enabled. Allow ingress from the proxy-only subnet and 136.124.104.0/22 in trust-vpc, and configure a custom static route with the proxy load balancer's forwarding rule as the next hop.This configuration attempts to deploy a regional internal proxy Network Load Balancer with GCE_VM_IP_PORT zonal NEGs as the next hop of a custom static route, permitting health checks from 136.124.104.0/22 and traffic from a proxy-only subnet.
136.124.104.0/22 range is used by global external passthrough Network Load Balancers, not internal load balancers. Internal load balancers send IPv4 health check probes from 35.191.0.0/16, so health checks in trust-vpc would be blocked by the implied deny ingress firewall rule.GCE_VM_IP_PORT endpoints for TCP proxy traffic.L3_DEFAULT or custom route next hop) with GCE_VM_IP zonal NEGs to preserve original client and destination IP addresses.The workload were a standard internal TCP application behind an Envoy proxy (rather than routed NVA traffic via a static route next hop), and the firewall rule allowed 35.191.0.0/16 instead of 136.124.104.0/22.
Always match the health check probe CIDR range to the specific load balancer type (35.191.0.0/16 for internal passthrough NLBs) and use GCE_VM_IP NEGs for next-hop routing.
GCE_VM_IP zonal NEGs targeting nic1 across multiple zones with HTTP_COOKIE session affinity and a regional UDP health check targeting the inspection port. Allow ingress from 35.191.0.0/16 in trust-vpc, verify that the Google guest agent creates the local route for the load balancer VIP, and rely on the VM's default routing table to return health check responses.This configuration correctly selects GCE_VM_IP zonal NEGs for nic1 and allows the 35.191.0.0/16 health check range, but attempts to use HTTP_COOKIE session affinity, a UDP health check, and the VM's default routing table for return traffic.
HTTP_COOKIE session affinity is a Layer 7 feature exclusive to Application Load Balancers. A Layer 4 internal passthrough Network Load Balancer cannot inspect HTTP headers or cookies and instead uses CLIENT_IP (or 3-tuple/5-tuple) hashing for session affinity.nic1: Relying on the VM's default routing table causes outbound health check replies (sourced from the load balancer VIP) to egress through nic0 (untrust-vpc) instead of nic1 (trust-vpc), causing probes to fail and marking all NVAs unhealthy.A TCP or HTTP health check were used instead of UDP, CLIENT_IP session affinity replaced HTTP_COOKIE, and policy-based routing (ip rule add from <LB_VIP>/32) were configured on each NVA for nic1.
Configuring failover backends and connection draining on failover also remains essential to avoid abrupt connection drops when an NVA fails.
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