Skip to content

Professional Cloud Network Engineer · Domain 5: Managing, monitoring, and troubleshooting network operations · 14% of the exam

Task 5.1: Logging and monitoring with Google Cloud Observability.

Logs for Cloud VPN, Cloud Router, VPC Service Controls, Cloud NGFW, VPC Flow Logs, Cloud DNS, Cloud NAT and Network Connectivity Center, and metrics for VPN, Cloud Interconnect, Cloud Router, load balancers, Cloud Armor and Cloud NAT.

Study it

Sample questions

Try each one before opening the answer. Every option is explained, with the Google Cloud documentation page that proves it.

Question 1 · choose 1

After a series of HA VPN tunnel flaps, the network team wants to compare Cloud VPN gateway log entries, such as IKE negotiation messages, across the past year whenever a tunnel misbehaves. Today the team reads them only in the Logs Explorer. What should it set up?

  1. AVPC Flow Logs on the VPN tunnels, kept in Cloud Logging for the team to query
  2. BA log sink that routes the Cloud VPN gateway's logs to a BigQuery dataset
  3. CA custom dashboard of the tunnels' Monitoring metrics, saved for a year
  4. DAn alerting policy on the tunnels' dropped packets that emails the team
Show the answer and why
  • AVPC Flow Logs on the VPN tunnels, kept in Cloud Logging for the team to query

    Incorrect

    Flow logs record sampled traffic flows through the tunnels, not the gateway's own log entries such as IKE messages.

  • BA log sink that routes the Cloud VPN gateway's logs to a BigQuery dataset

    Correct

    Cloud Logging keeps Cloud VPN logs for only 30 days; to keep them longer they must be routed, for example to BigQuery for analysis.

  • CA custom dashboard of the tunnels' Monitoring metrics, saved for a year

    Incorrect

    Metrics such as bytes and dropped packets describe tunnel traffic; they don't carry the gateway's log messages.

  • DAn alerting policy on the tunnels' dropped packets that emails the team

    Incorrect

    An alert notifies about a condition but stores none of the gateway's log entries for later comparison.

Cloud VPN gateways send logs to Cloud Logging and tunnels send metrics to Cloud Monitoring. The VPN documentation notes that Cloud Logging stores Cloud VPN logs for only 30 days, so a longer history needs a route to Pub/Sub or BigQuery.

Question 2 · choose 1

VMs without external IP addresses open TCP connections through a Public NAT gateway with default timeouts to a partner's message service. The partner pushes a message on each connection about every 30 minutes, and the VMs send nothing in between. Many messages never arrive. In Cloud Monitoring, nat/dropped_received_packets_count rises each time, nat/dropped_sent_packets_count stays at zero, and nat_allocation_failed is false. Which change should stop the losses without code changes?

  1. AAdd NAT IP addresses to the gateway so that more source ports are available to the VMs
  2. BRaise the TCP Transitory Connection Idle Timeout above 30 minutes
  3. CRaise the TCP Established Connection Idle Timeout above 30 minutes
  4. DTurn off endpoint-independent mapping so that each connection can use a different NAT address and port
Show the answer and why
  • AAdd NAT IP addresses to the gateway so that more source ports are available to the VMs

    Incorrect

    Missing addresses or ports drop sent packets with OUT_OF_RESOURCES and can set nat_allocation_failed; here only received packets are dropped.

  • BRaise the TCP Transitory Connection Idle Timeout above 30 minutes

    Incorrect

    The transitory timeout covers only half-open connections; these connections are established.

  • CRaise the TCP Established Connection Idle Timeout above 30 minutes

    Correct

    After the default 20 idle minutes the gateway removes the mapping, so the partner's next packet matches no connection entry and is dropped.

  • DTurn off endpoint-independent mapping so that each connection can use a different NAT address and port

    Incorrect

    Endpoint-independent mapping conflicts drop sent packets, with the reason ENDPOINT_INDEPENDENCE_CONFLICT; they don't remove idle mappings.

A Cloud NAT gateway drops an inbound packet when its connection tracking table has no entry for the connection, for example because the TCP Established Connection Idle Timeout, 1200 seconds (20 minutes) by default, expired. Connections that sit idle for 30 minutes therefore lose their mappings; a longer established idle timeout, or keepalives in the application, keeps them. Running out of NAT addresses or ports, and endpoint-independent mapping conflicts, show up as dropped sent packets.

Question 3 · choose 1

A security team wants a log entry for every DNS query that VMs in a VPC network send to the network's Cloud DNS private zones. How is this logging turned on?

  1. ABy enabling VPC Flow Logs on every subnet with metadata included
  2. BBy enabling Data Access audit logs for Compute Engine
  3. CBy turning on logging in each private zone's record sets
  4. DBy enabling logging in a DNS server policy attached to the network
Show the answer and why
  • ABy enabling VPC Flow Logs on every subnet with metadata included

    Incorrect

    Flow logs record sampled traffic flows, not each DNS query and answer.

  • BBy enabling Data Access audit logs for Compute Engine

    Incorrect

    Cloud DNS query logging is configured through DNS policies, not Compute Engine audit logs.

  • CBy turning on logging in each private zone's record sets

    Incorrect

    Query logging is set for networks, through a DNS policy, not per record set.

  • DBy enabling logging in a DNS server policy attached to the network

    Correct

    DNS policies enable or disable logging for networks; with it on, every query to a private managed zone is logged.

Cloud DNS query logging for private zones is controlled per VPC network with a DNS server policy: enable logging on the policy that applies to the network, or create one, and every DNS query to a private managed zone from that network is logged.

Question 4 · choose 1

An operations team wants one view of packets sent, packets received and dropped packets across all of a project's Cloud VPN tunnels, filterable by region, gateway and tunnel. Where should it look first?

  1. AThe Observability tab of the Cloud VPN page
  2. BThe Cloud Router's BGP session details
  3. CThe Logs Explorer, filtered to Cloud VPN gateway logs
  4. DFirewall Insights for the VPC network
Show the answer and why
  • AThe Observability tab of the Cloud VPN page

    Correct

    This predefined dashboard shows project-level metrics such as packets sent, packets received and dropped packets, with filters.

  • BThe Cloud Router's BGP session details

    Incorrect

    BGP session details show routing status, not tunnel packet metrics for the whole project.

  • CThe Logs Explorer, filtered to Cloud VPN gateway logs

    Incorrect

    Gateway logs hold log entries; packet counts are Cloud Monitoring metrics.

  • DFirewall Insights for the VPC network

    Incorrect

    Firewall Insights analyzes firewall rules, not VPN tunnel traffic.

Cloud VPN tunnels send metrics to Cloud Monitoring. The Observability tab on the Cloud VPN page is a predefined dashboard of project-level metrics, such as packets sent, packets received and dropped packets, filterable by region, gateway and tunnel; Metrics Explorer serves custom analysis.

Question 5 · choose 1

A team added a deny rule to a Cloud Armor security policy in preview mode and wants to see, for each backend service, how many requests the rule would have blocked, before it enforces the rule. Where can it see this?

  1. AIn VPC Flow Logs of the backend subnets
  2. BIn Cloud Armor's Cloud Monitoring metrics
  3. CIn the load balancer's backend health status
  4. DNowhere: preview mode discards its results
Show the answer and why
  • AIn VPC Flow Logs of the backend subnets

    Incorrect

    Flow logs don't know about Cloud Armor rule decisions.

  • BIn Cloud Armor's Cloud Monitoring metrics

    Correct

    Cloud Armor monitoring shows request volumes allowed, denied or previewed per policy and per backend service.

  • CIn the load balancer's backend health status

    Incorrect

    Health status reports whether backends pass health checks, not security policy decisions.

  • DNowhere: preview mode discards its results

    Incorrect

    Previewed requests are counted in the monitoring metrics.

Preview mode lets a Cloud Armor rule be evaluated without being enforced. Cloud Armor exports metrics to Cloud Monitoring that show request traffic allowed, denied or previewed per security policy and per backend service, so the effect of a previewed rule can be judged before enforcing it.

Practice domain 5 →Practice all domains →