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 team runs high-concurrency Cloud Run and App Engine standard services in us-central1 attached to a Shared VPC via an e2-micro Serverless VPC Access connector in the host project. Both services must:
During load testing with frequent small requests, Memorystore calls time out below 100 Mbps, and PSC traffic fails to route privately.
Which configuration resolves all issues?
e2-standard-4 instances on a dedicated, pre-created /28 PRIVATE-purpose subnet in the host project, attach Cloud NAT to that subnet, and set VPC egress to all-traffic on both services.e2-standard-4 connector with all-traffic egress isThis architecture pairs a high-capacity Serverless VPC Access connector using dedicated-core e2-standard-4 VMs and an explicit /28 PRIVATE-purpose subnet with full VPC egress routing (all-traffic) and Cloud NAT.
e2-micro to e2-standard-4 provides dedicated vCPUs and larger connection tracking memory, preventing packet-rate exhaustion during high-concurrency, small-request workloads.--vpc-egress=all-traffic on Cloud Run and egress_setting: all-traffic in App Engine's app.yaml forces traffic destined for privately used external IP addresses (such as the PSC endpoint for Google APIs) into the Shared VPC rather than the public internet./28 PRIVATE-purpose subnet in the Shared VPC host project makes the subnet visible for Cloud NAT static IP attachment. Keeping a connector avoids the 30-second or longer cold-start delays seen with Direct VPC egress and Cloud NAT while minimizing subnet IP consumption.all-traffic, outbound partner API calls obey VPC routes, VPC firewall rules, and Cloud NAT static IP translation.e2-standard-4 capacity with application-level connection pools prevents long-lived connection drops during biweekly 1-minute connector instance replacements.Unlike e2-micro connectors, Direct VPC egress, or hidden custom-range subnets, this design satisfies packet-rate performance, privately used external IP routing, explicit Cloud NAT integration, and Shared VPC IP conservation simultaneously.
e2-standard-4 instances using a custom /28 CIDR range, configure Cloud NAT on the automatically generated aet- subnet, and set VPC egress to all-traffic.When creating a Serverless VPC Access connector in the same project as its VPC network, supplying an unused /28 custom IP range instructs Google Cloud to automatically provision an internal subnet prefixed with aet- for the connector.
aet- subnet, outbound partner API traffic routed through the connector with all-traffic cannot use a dedicated static Cloud NAT IP address.aet- subnets only appear in the VPC network's subnet list when describing the network itself, preventing granular subnet-level administrative controls./28 PRIVATE-purpose subnet in the host project./28 subnet must remain dedicated solely to that single connector and cannot host VMs, load balancers, or PSC attachments.All outbound traffic went strictly to internal VPC or PSC destinations without requiring Cloud NAT on the connector's subnet, and no subnet-specific firewall targeting was needed.
--vpc-egress=all-traffic, and configure Cloud NAT with a static IP on that subnet.Direct VPC egress lets Cloud Run and App Engine standard services send outbound traffic directly to a VPC network by assigning IP addresses from a chosen subnet to individual serverless instances, eliminating connector VMs.
/28 connector pool, it consumes significantly more IP addresses in most high-concurrency deployments.The workload did not require Cloud NAT (or could tolerate 30+ second startup delays with startup probes) and the Shared VPC had ample IP address space for per-instance IP allocation.
While Direct VPC egress is recommended for most new serverless deployments, IP address scarcity and Cloud NAT startup sensitivity are the primary architectural reasons to retain Serverless VPC Access connectors.
/28 to /24, keep e2-micro instances with max-instances set to 10, and set VPC egress to private-ranges-only.e2-micro instances doesThis approach attempts to resolve throughput and timeout issues by enlarging the connector's IP address space, allowing more connector VM instances to scale out, and restricting VPC routing to private address blocks.
/28 primary IPv4 subnet. Once a connector is created, its subnet cannot be expanded to /24 or modified.e2-micro and f1-micro VMs, triggering timeouts well before bandwidth limits are reached. Increasing max-instances does not eliminate per-instance throttling under bursty small-packet loads.private-ranges-only routes only RFC 1918, RFC 6598, and 199.36.153.4/30 or 199.36.153.8/30 destinations into the VPC. Traffic to a PSC endpoint configured with a privately used external IP and outbound requests to external partner APIs bypass the connector and exit via the default shared internet pool.max-instances (maximum 10), but never scale back in, locking in VM compute costs.e2-micro instances rely on shared-core burst CPU credits that degrade rapidly under high packets-per-second workloads.private-ranges-only ignores custom VPC routes and Cloud NAT configurations for public or privately used external IP destinations.The connector were already using e2-standard-4 instances, no PSC endpoint with a privately used external IP or static Cloud NAT egress was required, and only RFC 1918 VPC destinations were accessed.
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