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 a new GKE Standard cluster to host a high-scale microservices platform with 4,000 ClusterIP Services and 120,000 backend Pod endpoints. Previous clusters using traditional iptables-based kube-proxy routing experienced latency spikes and slow rule programming as the Service count grew.
You must provision and verify the new cluster to meet the following requirements:
kube-proxy and iptables Service routing with kernel-space eBPF packet processing that leverages Kubernetes metadata.NetworkPolicy natively without third-party CNI add-ons or conflicting configuration flags.Which provisioning and verification workflow should you implement?
--enable-network-policy, run gcloud container clusters update with --enable-dataplane-v2 to convert kube-proxy from iptables mode to ipvs mode, and verify networkConfig.datapathProvider returns ADVANCED_DATAPATH.kube-proxy ipvs mode areRoutes-based clusters use Google Cloud VPC static routes rather than VPC alias IP ranges (VPC-native routing) to route Pod IP addresses across nodes. Meanwhile, ipvs mode is a Linux IP Virtual Server operating mode for kube-proxy used in clusters without GKE Dataplane V2 to replace iptables Service rules with kernel hash-table load balancing.
--enable-network-policy) cannot be migrated directly without first removing policies and disabling Calico.kube-proxy into ipvs mode. Instead, Dataplane V2 eliminates kube-proxy altogether and performs Service ClusterIP load balancing and DNAT through Cilium eBPF programs managed by anetd.networkConfig.datapathProvider omits verifying the anetd DaemonSet on worker nodes.No valid scenario supports this exact workflow, because routes-based clusters cannot migrate to Dataplane V2 and Dataplane V2 never runs kube-proxy in ipvs mode.
gcloud container clusters create with --enable-dataplane-v2 (without --enable-network-policy), verify networkConfig.datapathProvider returns ADVANCED_DATAPATH via gcloud container clusters describe, and confirm anetd Pods with label k8s-app=cilium are running in kube-system.GKE Dataplane V2 is a Kubernetes-optimized networking dataplane powered by Cilium and eBPF (Extended Berkeley Packet Filter). Instead of relying on kube-proxy and linear iptables rule chains to handle Service virtual IP (VIP) translation, GKE Dataplane V2 executes eBPF programs directly inside the Linux kernel and uses Kubernetes-specific metadata to perform routing, load balancing, and policy enforcement in kernel space.
kube-proxy and iptables: Passing --enable-dataplane-v2 during gcloud container clusters create deploys the anetd DaemonSet to each node. anetd watches Kubernetes objects and programs eBPF Service maps directly in the kernel, eliminating iptables chain traversal bottlenecks. With 4,000 ClusterIP Services and 120,000 backend endpoints, the cluster stays well within the 10,000 ClusterIP Service and 260,000 eBPF Service map endpoint limits.NetworkPolicy enforcement without conflicts: GKE Dataplane V2 always enables Kubernetes NetworkPolicy enforcement natively via Cilium. Omitting --enable-network-policy is critical because attempting to explicitly enable or disable legacy network policy enforcement on a cluster with DatapathProvider=ADVANCED_DATAPATH causes the request to fail.gcloud container clusters describe formatted for networkConfig.datapathProvider confirms ADVANCED_DATAPATH at the control plane, while querying kube-system Pods matching k8s-app=cilium confirms the anetd- controller Pods are active on every node.kube-proxy CPU and latency overhead caused by updating massive iptables chains.NetworkLogging CRD for auditing allowed and denied connections and integrates with Hubble.Unlike legacy Calico or kube-proxy configurations, enabling --enable-dataplane-v2 at cluster creation provides an integrated, high-scale eBPF architecture out of the box without conflicting flags, unsupported custom eBPF programs, or ineligible cluster migrations.
gcloud container clusters create with --enable-dataplane-v2, deploy a custom eBPF DaemonSet on each node to bypass the 100,000-endpoint eBPF map ceiling, and confirm kube-proxy Pods are replaced by hubble-relay in kube-system.Custom eBPF programs are user-deployed kernel hooks or third-party observability/networking agents attached directly to node network interfaces. Hubble Relay (hubble-relay) is the telemetry aggregation service deployed when GKE Dataplane V2 flow observability is enabled, allowing Hubble CLI and Hubble UI to inspect live cluster traffic flows.
anetd (in the kube-system namespace) is the eBPF networking controller that replaces kube-proxy. By contrast, hubble-relay is an optional observability component deployed in the gke-managed-dpv2-observability namespace only when --enable-dataplane-v2-flow-observability is enabled.anetd reported memory metrics.hubble-relay) is independent of core Dataplane V2 packet forwarding.No supported GKE configuration permits custom eBPF programs alongside Dataplane V2; for 120,000 endpoints, native anetd eBPF maps handle routing without custom kernel extensions.
gcloud container clusters create with both --enable-dataplane-v2 and --enable-network-policy, verify networkConfig.datapathProvider returns ADVANCED_DATAPATH, and confirm calico-node Pods are managing eBPF enforcement alongside kube-proxy.--enable-network-policy and Calico doThe --enable-network-policy flag enables legacy Kubernetes NetworkPolicy enforcement on GKE Standard clusters by deploying the Calico Container Network Interface (CNI) add-on (calico-node DaemonSets). In the legacy datapath architecture, Calico programs iptables rules in the Linux kernel for Pod policy enforcement while kube-proxy separately manages iptables or IPVS rules for Service routing.
gcloud flags: When --enable-dataplane-v2 is specified, Kubernetes NetworkPolicy enforcement is built in and permanently active. Attempting to combine --enable-dataplane-v2 with --enable-network-policy—or later toggling network policy enforcement on a cluster with DatapathProvider=ADVANCED_DATAPATH—is explicitly disallowed and causes GKE to reject the operation with an error.anetd Pods in kube-system rather than calico-node Pods. Furthermore, Dataplane V2 replaces kube-proxy rather than running alongside it.iptables dependency: Calico's legacy GKE integration relies on iptables rules, which reintroduces rule-evaluation overhead at scale.kube-proxy splits policy enforcement and Service routing across two independent node agents.NetworkLogging CRD or Hubble-based flow observability.You were intentionally provisioning a legacy Dataplane V1 cluster (LEGACY_DATAPATH) without --enable-dataplane-v2 and needed Calico NetworkPolicy enforcement managed alongside kube-proxy.
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