Prepare and test your skills

Prepare and test your skills

Free sample question with answers and detailed explanation below. The Professional Cloud Developer practice bank has an additional 608 practice questions.Try 10 free realistic questions: instant score, no credit card, and no account required.
You are deploying a multi-service web application onto a VPC-native Google Kubernetes Engine (GKE) cluster. The application consists of separate frontend and API microservices that experience unpredictable traffic spikes throughout the day.
Your deployment must meet the following requirements:
/api and /web).Which configuration should you implement?
Combining resource requests in Deployment manifests, a HorizontalPodAutoscaler (HPA), and a GKE Ingress with container-native load balancing provides end-to-end workload scaling and Layer 7 traffic management on GKE.
requests for CPU in your container specifications gives the Kubernetes scheduler accurate baseline requirements and allows Metrics Server to compute CPU utilization percentages. The HPA continuously queries these metrics and automatically adds or removes Pod replicas as load fluctuates./api and /web to distinct backend Services.requests prevent HPA calculation errors and ensure reliable scheduling under load.Unlike Layer 4 LoadBalancer Services or node-based NodePort backends, HPA paired with Ingress and NEGs delivers elastic horizontal Pod scaling alongside efficient, resilient Layer 7 routing.
VerticalPodAutoscaler (VPA) dynamically adjusts the CPU and memory requests and limits of existing Pods, while a Service of type: LoadBalancer provisions a regional external passthrough Network Load Balancer at Layer 4. Setting externalTrafficPolicy: Local instructs the load balancer to send traffic only to nodes running healthy backend Pods.
LoadBalancer Service operates at Layer 4 (TCP/UDP) for a single Service and cannot inspect HTTP paths such as /api or /web to route traffic across multiple Services.externalTrafficPolicy: Local avoids extra cross-node hops, it still relies on node-level health checks rather than Pod-direct NEGs and can cause uneven traffic distribution if Pods are not balanced evenly across nodes.type: LoadBalancer provisions separate external IP addresses and load balancers.You were deploying a single non-HTTP TCP or UDP service that could not scale horizontally and needed vertical resource sizing with node-local traffic preservation.
Cluster Autoscaler monitors cluster capacity and automatically adds or removes worker nodes in node pools based on Pod scheduling needs. Exposing Services behind an Ingress without NEGs uses instance groups (NodePort backends) where the load balancer routes traffic to cluster nodes first, and kube-proxy forwards packets to backend Pods.
NodePort forwarding across nodes rather than targeting Pod IP addresses directly.You only needed to resize node pools for manually scaled workloads in a legacy routes-based cluster that does not support VPC-native container-native load balancing.
This configuration correctly uses CPU requests and HPA for horizontal Pod scaling, but attempts to expose multiple microservices and route by URI path using a Service of type: LoadBalancer paired with loadBalancerSourceRanges.
loadBalancerSourceRanges: The loadBalancerSourceRanges field in a Service manifest specifies a list of allowed client CIDR IP address ranges to restrict network access at the node firewall level. It performs Layer 3/4 IP allowlisting and cannot inspect or route HTTP URI paths such as /api or /web.LoadBalancer Service provisions a Layer 4 Network Load Balancer bound to a single Service selector, meaning it cannot perform HTTPS path-based routing across separate frontend and API Services.LoadBalancer Services route to node VMs via kube-proxy or eBPF rather than using NEGs to target Pod IP addresses directly.loadBalancerSourceRanges acts strictly as an access control filter for client source IPs.LoadBalancer Service forwards traffic to one set of matching Pods rather than acting as a multi-Service HTTP router.You needed to expose a single TCP or UDP Service externally while restricting access to a specific set of corporate or partner client IP CIDR blocks.
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