Try each one before opening the answer. Every option is explained, with the Google Cloud documentation page that proves it.
Question 1 · choose 1
A Cloud Armor backend security policy on a global external Application Load Balancer's backend service has these rules: priority 100 allows 198.51.100.0/24, priority 1000 denies 198.51.100.64/26 with a 403, and the default rule allows everything else. Requests from 198.51.100.70 are still allowed. The rest of 198.51.100.0/24 must stay allowed. What should the network engineer change?
AChange the deny rule's priority from 1000 to 50,000
BChange the default rule's action from allow to deny with a 403
CAttach a second backend security policy that holds only the deny rule to the same backend service
DChange the deny rule's priority from 1000 to 50
Show the answer and why
AChange the deny rule's priority from 1000 to 50,000
Incorrect
A larger number means a lower priority, so the allow rule at 100 would still match first.
BChange the default rule's action from allow to deny with a 403
Incorrect
Requests from 198.51.100.70 match the allow rule at 100 before they ever reach the default rule; other clients would be blocked instead.
CAttach a second backend security policy that holds only the deny rule to the same backend service
Incorrect
A backend service can't have two backend security policies at the same time.
DChange the deny rule's priority from 1000 to 50
Correct
Rules are evaluated from the lowest priority number up, so a deny at 50 matches the /26 before the allow at 100 matches the rest of the /24.
Cloud Armor evaluates a security policy's rules in priority order, lowest number first, and stops at the first match; the default rule has the highest possible number and applies only when nothing else matches. A narrower deny must come before the broader allow that contains it.
A team is adding Cloud Armor's preconfigured SQL injection rules to a busy production policy and wants to limit the risk of blocking legitimate traffic while it learns how the rules behave. How should it start, following Google's guidance?
AevaluatePreconfiguredWaf('sqli-v422-stable', {'sensitivity': 1}) in preview mode
BevaluatePreconfiguredWaf('sqli-v422-stable', {'sensitivity': 4}) enforced at once
CevaluatePreconfiguredWaf('sqli-v422-stable', {'sensitivity': 1}) enforced at once
DA custom rule that denies every request containing a single quote
Show the answer and why
AevaluatePreconfiguredWaf('sqli-v422-stable', {'sensitivity': 1}) in preview mode
Correct
Google recommends starting with sensitivity level 1 in preview mode, then tuning before enforcing.
BevaluatePreconfiguredWaf('sqli-v422-stable', {'sensitivity': 4}) enforced at once
Incorrect
The highest sensitivity enforced immediately carries the most risk of false positives.
CevaluatePreconfiguredWaf('sqli-v422-stable', {'sensitivity': 1}) enforced at once
Incorrect
The lowest sensitivity is the right start, but enforcing at once skips the preview trial that the guidance asks for.
DA custom rule that denies every request containing a single quote
Incorrect
A blanket pattern would block legitimate requests; the preconfigured rules exist to avoid hand-written signatures.
Preconfigured WAF rules are evaluated with evaluatePreconfiguredWaf() and a sensitivity level. Cloud Armor's guidance is to start with sensitivity level 1 in preview mode, which logs matches without enforcing them, and to tune the rules before raising sensitivity or enforcing.
A Cloud Armor security policy has a rule with the rate_based_ban action that bans clients for 10 minutes once they exceed 300 requests per minute. The product team now wants clients above the threshold only throttled, and no client may ever be banned. What should the network engineer do?
AEdit the existing rule and change its action from rate_based_ban to throttle
BKeep the rule and turn on preview mode, so that it throttles clients instead of banning them
CCreate a new rule with the throttle action and remove the rate-based ban rule
DKeep the rule and change its exceed action from deny to redirect, which ends the bans
Show the answer and why
AEdit the existing rule and change its action from rate_based_ban to throttle
Incorrect
A rule configured with the rate-based ban action can't be changed to a throttle action later.
BKeep the rule and turn on preview mode, so that it throttles clients instead of banning them
Incorrect
Preview mode shows a rule's effects in the request logs; it doesn't throttle anything.
CCreate a new rule with the throttle action and remove the rate-based ban rule
Correct
A ban rule can't become a throttle rule, so a new throttle rule must replace it.
DKeep the rule and change its exceed action from deny to redirect, which ends the bans
Incorrect
A rate-based ban rule applies its exceed action to all of a client's requests for the ban duration, whichever action it is.
A throttle rule can later be changed to a rate-based ban rule, but a rule configured with the rate-based ban action can't be changed to a throttle action. To stop banning clients, the team creates a throttle rule with the wanted threshold and removes the rate-based ban rule.
An existing Cloud Armor policy allows requests only from a third-party CDN by referencing a Cloud Armor named IP address list maintained by that provider. A new team member asks whether to keep building rules this way. What should the network engineer answer?
AYes: named IP address lists are the current way to reference provider ranges
BYes, but only in the Standard tier, where named lists are free
CNo: copy the ranges into the rule's expression once and never update them
DNo: they're deprecated; Google recommends Threat Intelligence feeds
Show the answer and why
AYes: named IP address lists are the current way to reference provider ranges
Incorrect
Named IP address lists are deprecated on the Threat Intelligence page.
BYes, but only in the Standard tier, where named lists are free
Incorrect
The deprecation applies regardless of tier.
CNo: copy the ranges into the rule's expression once and never update them
Incorrect
Provider ranges change; a frozen copy isn't the recommended replacement.
DNo: they're deprecated; Google recommends Threat Intelligence feeds
Correct
The Cloud Armor documentation marks named IP address lists as deprecated and recommends Threat Intelligence feeds.
Cloud Armor named IP address lists let rules reference IP ranges maintained by third-party providers, but they are deprecated. The documentation recommends Google Threat Intelligence feeds, a Cloud Armor Enterprise feature, instead.