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.
A GKE Standard cluster uses Cloud DNS in VPC scope with default-deny NetworkPolicy rules. After another Cloud DNS response policy is bound to the cluster's VPC network, you observe:
FailedPrecondition (Failed to send requests to Cloud DNS due to a previous error).api.vendor.com (with nameserver 169.254.169.254) experience high latency and NXDOMAIN query bursts for api.vendor.com.default.svc.cluster.local.kubedns latency metrics remain flat.Which actions should you take to restore DNS synchronization and reduce external DNS latency?
This option attempts to tune the in-cluster kube-dns Deployment and kube-dns ConfigMap while adding a VPC firewall rule for traffic targeting the local metadata server IP (169.254.169.254).
nameserver 169.254.169.254 in /etc/resolv.conf, confirming they send DNS queries directly to the node-local Cloud DNS data plane, completely bypassing kube-dns Pods. Scaling kube-dns has zero effect on query latency or capacity, which also explains why kubedns metrics are flat.kube-dns ConfigMap (stubDomains and upstreamNameservers). Custom forwarding or upstream settings must be configured directly in Cloud DNS forwarding zones or server policies.169.254.169.254 on the node, nor do they stop ndots:5 search-path NXDOMAIN bursts.kube-dns only handles DNS when Cloud DNS for GKE is disabled or on un-migrated node pools.FailedPrecondition synchronization failures stem from VPC-level Cloud DNS resource conflicts or API quotas, not Pod-to-kube-dns capacity.The cluster were running kube-dns instead of Cloud DNS for GKE, experiencing concurrent query rejections on kube-dns Pods, and needing upstream forwarding to resolve external domains without VPC firewall blocks between nodes and kube-dns Pods.
This option tries to bypass the GKE Cloud DNS controller by manually inserting DNS records, performing a same-version node pool upgrade, and requesting a Cloud DNS API write quota increase.
gke-...-dns) and response policy (gke-...-rp). Manual changes to records in these zones are unsupported and are overwritten by the Cloud DNS controller./etc/resolv.conf already shows 169.254.169.254, proving Pods are already using Cloud DNS.FailedPrecondition trigger: Although InsufficientQuota (Write requests per minute) can trigger cascading FailedPrecondition errors, the scenario explicitly identifies another Cloud DNS response policy bound to the VPC network as the preceding event.ndots:5 latency: External NXDOMAIN query amplification (api.vendor.com.default.svc.cluster.local) and missing node-local caching remain unresolved.kube-dns to Cloud DNS when upgrading to a new GKE version or creating a new node pool.The cluster events showed Warning InsufficientQuota on the managedzone resource due to exceeding the regional dns.googleapis.com write-requests limit, and legacy node pools were still pointing /etc/resolv.conf to a 10.x.y.10 kube-dns Service IP.
This option proposes an in-place Cloud DNS scope change to additive VPC scope, creates a VPC ingress firewall rule for 35.199.192.0/19, and restarts kube-dns Pods.
--cluster-dns-scope is configured on a cluster, switching between VPC scope and cluster scope (or changing scope) cannot be performed in place; the cluster must be recreated.35.199.192.0/19 range: Google Cloud uses 35.199.192.0/19 as the source IP range when Cloud DNS outbound forwarding sends queries to external or on-premises DNS servers. Creating a VPC ingress firewall rule for this range does not fix GKE-to-Cloud DNS record synchronization or Pod external lookup latency.kube-dns is irrelevant: Pods are using 169.254.169.254 (Cloud DNS), not kube-dns, and ndots:5 search-path traversal is controlled by each Pod's /etc/resolv.conf configuration rather than server-side caching.35.199.192.0/19 firewall rules belong on target on-premises or remote DNS servers receiving forwarded Cloud DNS queries (unless using Cloud DNS outbound endpoints with a proxy-only subnet).On-premises DNS servers were dropping outbound forwarded queries from Cloud DNS forwarding zones because 35.199.192.0/19 on TCP/UDP port 53 was blocked by an on-premises firewall.
This approach resolves the underlying Cloud DNS response policy conflict halting control-plane synchronization while optimizing the Pod DNS resolution path through local node caching and reduced search-path amplification.
gke-...-rp) to the VPC for Service PTR records. Deleting the conflicting custom response policy clears the FailedPrecondition blocker so the Cloud DNS controller can resume syncing new Service records.options ndots:5, causing external names like api.vendor.com (1 dot) to traverse all five search suffixes first, generating five NXDOMAIN lookups before the root query. Setting ndots:1 or appending a trailing dot (api.vendor.com.) bypasses unnecessary suffix traversal.169.254.20.10), reducing upstream calls to Cloud DNS via 169.254.169.254. Adding an egress NetworkPolicy for TCP and UDP port 53 ensures default-deny policies do not drop queries to node-local-dns.kube-dns Pods, kubedns metrics remain flat; evaluating Cloud DNS query/latencies and query/response_count (alongside node_local_dns metrics) provides accurate visibility.FailedPrecondition errors across all cluster Services without recreating the cluster.NXDOMAIN search-path storms.NetworkPolicy enforcement while permitting required port 53 DNS flows.Unlike modifying kube-dns or manually editing managed zones—both ineffective in Cloud DNS VPC scope—removing the conflicting VPC response policy directly fixes the control-plane failure while NodeLocal DNSCache and ndots:1 address the exact source of external lookup latency.
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