Try each one before opening the answer. Every option is explained, with the Google Cloud documentation page that proves it.
Question 1 · choose 1
Security deleted the default 0.0.0.0/0 route of a VPC network so that its VMs cannot reach the internet. VMs without external IP addresses in a subnet with Private Google Access enabled must still reach Cloud Storage, and private DNS already resolves storage.googleapis.com to the private.googleapis.com range 199.36.153.8/30. Those requests now time out. What should you add?
AA Cloud NAT gateway on a Cloud Router in the subnet's region
BA custom static route for 199.36.153.8/30 with the default internet gateway as next hop
CAn ingress firewall rule allowing TCP 443 from 199.36.153.8/30 to the VMs
DA custom static route for 199.36.153.8/30 with one of the subnet's VMs as next hop
Show the answer and why
AA Cloud NAT gateway on a Cloud Router in the subnet's region
Incorrect
Cloud NAT provides outbound internet access, which security removed on purpose, and traffic to Google APIs is handled by Private Google Access rather than by NAT.
BA custom static route for 199.36.153.8/30 with the default internet gateway as next hop
Correct
Private Google Access needs routes to the Google API ranges whose next hop is the default internet gateway; a specific route replaces the deleted default route for this range only.
CAn ingress firewall rule allowing TCP 443 from 199.36.153.8/30 to the VMs
Incorrect
Firewall rules are stateful, so replies to the VMs' requests are already allowed; what is missing is a route for the outgoing requests.
DA custom static route for 199.36.153.8/30 with one of the subnet's VMs as next hop
Incorrect
Traffic to Google APIs cannot be routed through other VM instances or custom next hops; the route's next hop must be the default internet gateway.
A custom static route can be narrower than the default route. Here it sends only the private.googleapis.com range to the default internet gateway, so Google APIs work while the rest of the internet stays unreachable.
VMs in a private subnet have no external IP addresses and must stay unreachable from the internet. They now need to download OS packages from public mirrors. What should you configure?
AA Cloud NAT gateway on a Cloud Router for the subnet's region
BAn ephemeral external IP address on each VM
CPrivate Google Access on the subnet
DAn egress firewall rule allowing TCP 80 and 443 to 0.0.0.0/0
Show the answer and why
AA Cloud NAT gateway on a Cloud Router for the subnet's region
Correct
Cloud NAT lets VMs without external IPs open outbound connections to the internet, while unsolicited inbound connections are not allowed.
BAn ephemeral external IP address on each VM
Incorrect
External IPs would make the VMs reachable from the internet, which is not allowed.
CPrivate Google Access on the subnet
Incorrect
Private Google Access reaches Google APIs and services, not public package mirrors elsewhere on the internet.
DAn egress firewall rule allowing TCP 80 and 443 to 0.0.0.0/0
Incorrect
Egress is already allowed by the implied rule; the missing piece is a public source address for the traffic, which NAT provides.
Firewall rules decide what is allowed; addressing decides what is reachable. Cloud NAT gives private VMs outbound internet access without giving them public addresses.
A partner has allowlisted the current ephemeral external IP address of one of your VMs. You must keep exactly this address even if the VM is stopped or recreated. What should you do?
AReserve a new static external IP address and assign it to the VM
BConfigure Cloud NAT with the VM's current address as the NAT IP
CEnable deletion protection on the VM so that it keeps its address
DPromote the ephemeral external IP address to a static address
Show the answer and why
AReserve a new static external IP address and assign it to the VM
Incorrect
A new reservation gets a different address, so the partner's allowlist would no longer match.
BConfigure Cloud NAT with the VM's current address as the NAT IP
Incorrect
Cloud NAT uses addresses reserved for NAT and serves VMs without external IPs; it cannot keep an ephemeral VM address.
CEnable deletion protection on the VM so that it keeps its address
Incorrect
Deletion protection blocks deleting the VM; stopping it still releases an ephemeral address.
DPromote the ephemeral external IP address to a static address
Correct
Promoting reserves the address the VM already uses, so it stays in the project until you release it.
Ephemeral external addresses can change when a VM stops. If one must be kept, promote it to static in place rather than reserving another one.
A network has 150 VPC firewall rules. The team wants to move them into a global network firewall policy to use features such as FQDN objects, without rewriting every rule by hand. What should it use?
ACloud Asset Inventory export of the rules to BigQuery
BRaising every rule's priority to 1
CThe VPC firewall rules migration tool in the gcloud CLI
DA script that deletes all rules and asks teams to recreate them
Show the answer and why
ACloud Asset Inventory export of the rules to BigQuery
Incorrect
An export lists the rules; it does not convert them.
BRaising every rule's priority to 1
Incorrect
Priorities do not move rules into a policy.
CThe VPC firewall rules migration tool in the gcloud CLI
Correct
The tool creates a global network firewall policy and converts the existing rules, keeping their relative order.
DA script that deletes all rules and asks teams to recreate them
Incorrect
Recreating 150 rules by hand is slow and error-prone.
Network firewall policies add capabilities that VPC firewall rules lack, and the migration tool converts existing rules into one.
Security wants DNS answers for the company's public domain, hosted in a Cloud DNS public zone, to be protected against spoofing and cache poisoning. What should you enable?
ADNSSEC on the public zone, plus a DS record
BA private zone with the same records
CA shorter TTL on every record
DA firewall rule allowing only port 53
Show the answer and why
ADNSSEC on the public zone, plus a DS record
Correct
DNSSEC authenticates DNS responses; the registrar's DS record completes the chain of trust.
BA private zone with the same records
Incorrect
Moving the records out of the public zone would break public resolution of the domain.
CA shorter TTL on every record
Incorrect
TTLs control caching time, not response authenticity.
DA firewall rule allowing only port 53
Incorrect
Cloud DNS public zones are not protected by VPC firewall rules.
DNSSEC protects the integrity of public DNS answers; it requires work at both the zone and the registrar.
You want a new Cloud SQL instance to have only a private IP address in your VPC network. Creating it with private IP fails because no private connection exists. What should you set up first?
AA static external IP address for the instance
BPrivate services access with an allocated IP range
CA firewall rule allowing port 3306 from anywhere
DCloud NAT on the instance's region
Show the answer and why
AA static external IP address for the instance
Incorrect
This is the public address the design wants to avoid.
BPrivate services access with an allocated IP range
Correct
Cloud SQL private IP uses private services access, which needs an allocated range and a private connection.
CA firewall rule allowing port 3306 from anywhere
Incorrect
Firewall rules do not create the private connection.
DCloud NAT on the instance's region
Incorrect
This does not create a private connection to Cloud SQL.
Managed services such as Cloud SQL reach your VPC privately through private services access, which you set up once per network.
A VPC network has two custom static routes: 10.20.0.0/16 with priority 100 and 10.20.5.0/24 with priority 1000, with different next hops. A VM sends a packet to 10.20.5.7. Which route is used?
AThe 10.20.0.0/16 route, because it has the lower priority number
BNeither, because the default internet route wins
CBoth routes, with traffic split between them
DThe 10.20.5.0/24 route, because it is the most specific match
Show the answer and why
AThe 10.20.0.0/16 route, because it has the lower priority number
Incorrect
Priority is considered only among routes with the same, most specific destination.
BNeither, because the default internet route wins
Incorrect
The default route is the least specific and is used only when nothing more specific matches.
CBoth routes, with traffic split between them
Incorrect
Traffic is not split across routes with different prefixes.
DThe 10.20.5.0/24 route, because it is the most specific match
Correct
Routing picks the most specific destination first; priority only breaks ties between equally specific routes.
Route selection checks specificity before priority, which is why a narrow route can override a broad one regardless of priority.
A project still uses its default network, and an audit flags SSH and RDP open to the whole internet. Which rules in the default network should you delete or restrict first?
Adefault-allow-ssh and default-allow-rdp
BThe implied allow egress rule of the network
Cdefault-allow-internal in the default network
DThe implied deny ingress rule of the network
Show the answer and why
Adefault-allow-ssh and default-allow-rdp
Correct
These pre-populated rules allow SSH and RDP from any source.
BThe implied allow egress rule of the network
Incorrect
Egress rules do not open inbound SSH or RDP.
Cdefault-allow-internal in the default network
Incorrect
This rule allows traffic between instances inside the network, not from the internet.
DThe implied deny ingress rule of the network
Incorrect
This rule blocks unmatched inbound traffic; removing it is not possible and would not help.
The default network is convenient but permissive. Review its pre-populated rules before using it for anything real.