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 runs a GKE Standard cluster in a Shared VPC service project (svc-prod), with subnets located in the host project (vpc-host). VPC Flow Logs are enabled on all subnets (flow_sampling = 0.5, default metadata), and Observability Analytics is enabled on _Default in vpc-host. When you open Flow Analyzer in svc-prod, no flow data appears.
You must restore GKE visibility and analyze traffic to meet these goals:
Which approach should you take?
vpc-host after enabling INCLUDE_ALL_METADATA and Intranode Visibility, set Sampling points to Avg of source and destination, compare zone pairs or use Log Analytics for != zone filtering, inspect Egress AS paths in flow details, and use Latency mode with Max ranking and a shorter alignment period.Flow Analyzer in Network Intelligence Center provides visual, 5-tuple network traffic analysis over VPC Flow Logs stored in Observability Analytics-enabled Cloud Logging buckets. In a Shared VPC topology, subnet flow logs reside in the host project (vpc-host) rather than the GKE service project (svc-prod), making the host project the required entry point for querying and cost/performance diagnostics.
vpc-host accesses the subnet flow logs. Configuring VPC Flow Logs with metadata="INCLUDE_ALL_METADATA" populates GKE pod, workload, zone, and routing annotations, while enabling Intranode Visibility on the GKE Standard cluster hairpins same-node pod traffic through the VPC so it is captured in flow logs.!= operator, you either organize flows by source and destination zone to compare distinct zone pairs or click View in Log Analytics to run SQL filtering where src_instance.zone != dest_instance.zone.contains ASN or has first ASN).This approach resolves the Shared VPC project boundary issue, preserves full GKE workload and intra-node telemetry, prevents 2x overcounting of east-west traffic, works within Flow Analyzer's exact-match filter model, and tunes latency visualization specifically to expose short-lived anomalies.
vpc-host after enabling INCLUDE_ALL_METADATA, set Sampling points to Source endpoint to prevent double-counting, apply a not equal zone filter in the Flow Analyzer UI, and use Latency mode with Median ranking and a 1-day alignment period to reveal spikes.This approach correctly moves analysis to the Shared VPC host project and enables full metadata, but it relies on unsupported filter syntax, incomplete endpoint sampling for bidirectional visibility, and aggregation settings that mask transient latency anomalies.
OR logic within a field and AND logic across fields; they do not support a not equal (!=) operator to exclude intra-zone traffic.Instead of using an unsupported UI not equal filter, Median ranking, and a 1-day alignment period, you compared zone pairs in the results table or used Log Analytics SQL for != filtering, selected Avg of source and destination sampling, enabled Intranode Visibility, and switched to Max or P99 ranking with a shorter alignment period.
When troubleshooting latency in Flow Analyzer, Chart aggregation and Rank flows by work together: you can rank by Median to find consistently slow flows, or rank by Max or P99 with a fine-grained alignment period (such as 1 minute or 5 seconds for short windows) to isolate transient spikes.
svc-prod after upgrading its _Default bucket to Observability Analytics, set Sampling points to Sum of source and destination, query cross-zone flows in Log Analytics, and check Latency mode with P99 ranking and the Egress AS paths tab in flow details.This approach attempts to perform traffic analysis from the Shared VPC service project while summing flow logs reported by both sending and receiving endpoints, alongside valid techniques such as Log Analytics SQL querying, P99 latency ranking, and Egress AS path inspection.
vpc-host), not the service project (svc-prod) running the GKE workloads. Upgrading _Default in svc-prod will not surface subnet flow logs, so Flow Analyzer in svc-prod will continue showing no data. Additionally, GKE metadata requires INCLUDE_ALL_METADATA, and Standard clusters require Intranode Visibility for local pod-to-pod flows.!= zone comparisons, provided the query targets the host project's log bucket.Flow Analyzer were opened in vpc-host (or a centrally routed logging project), INCLUDE_ALL_METADATA and Intranode Visibility were enabled, and Sampling points were changed to Avg of source and destination to prevent double-counting.
For VM-to-VM or GKE node-to-node communication within subnets that have VPC Flow Logs enabled, both endpoints generate a flow record for the same connection. Flow Analyzer's Avg of source and destination option automatically divides dual-reported traffic by two within each alignment period.
vpc-host, raise flow_sampling to 1.0 on all subnets to recover missed historical flows, set Sampling points to Destination endpoint to deduplicate traffic, use Packet drops mode to measure TCP round-trip time spikes, and filter internet paths using Client - Server aggregation.This approach correctly identifies vpc-host as the project holding Shared VPC flow logs, but it relies on several misconceptions regarding flow sampling behavior, endpoint reporting, display modes, and connection aggregation.
flow_sampling from 0.5 (LIGHT tier) to 1.0 (COMPREHENSIVE tier) only affects future flow log collection; it cannot retroactively recover historical flows that were never sampled.gke_service definitions and aggregates bidirectional traffic, but inspecting autonomous system hops requires the Egress AS paths tab in Flow details and Egress AS path flow parameters.flow_sampling = 0.5 is already the recommended baseline for troubleshooting and cost analysis; 1.0 increases Cloud Logging ingestion costs for future traffic only.You kept flow_sampling at 0.5 (while enabling INCLUDE_ALL_METADATA and Intranode Visibility), used Avg of source and destination sampling, analyzed TCP RTT spikes in Latency mode with Max ranking, and viewed external routing in the Egress AS paths tab.
Even at sampling rates below 1.0 (such as 0.5 or 0.1), VPC Flow Logs estimates total byte and packet volumes from sampled packets, making 0.5 effective for cost optimization and top-talker analysis without incurring full 1.0 ingestion charges.
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