Question 950
A public service uses an ALB behind Global Accelerator. A regional endpoint's application health is degraded but its data-plane listener is still open. Which configuration improves traffic steering accuracy?
- Configure the target group's deregistration delay
- Attach multiple compatible certificates to the ALB listener's certificate list for SNI selection
- Select a supported ALB listener security policy that excludes disallowed legacy TLS protocols
- Ensure endpoint health checks reflect the application's readiness rather than mere port reachability
Explanation
Correct Answer
D:Health-based routing is only as meaningful as the endpoint health signal; a basic open TCP listener can remain available while application dependencies fail.
Incorrect Answers
A:The deregistration delay sets how long a deregistering target can continue draining existing requests before being fully removed from service; those conditions do not produce the required result: fail away from functionally unhealthy regions.
B:Server Name Indication lets the load balancer choose the appropriate configured certificate for the requested hostname without requiring one public IP per certificate; it changes a different aspect of the architecture rather than fulfilling the need to fail away from functionally unhealthy regions.
C:ALB security policies govern the protocols and ciphers accepted by its TLS listener; rotating the certificate alone does not disable old protocol versions; the operational gap would remain because the question requires a way to fail away from functionally unhealthy regions.
Question 951
A company receives unusually high hourly charges for interface endpoints in every subnet despite low API call volumes. Which design question should be evaluated first?
- Compare endpoint hourly and data-processing charges against actual traffic, AZ placement, and alternatives
- Add a DynamoDB gateway endpoint to the application subnet route tables
- Create an S3 gateway endpoint and associate the data subnet route tables
- Enable private DNS where supported and place endpoint ENIs in the active workload Availability Zones
Explanation
Correct Answer
A:Interface endpoint costs depend on deployed endpoint-AZ hours and data processed; low-volume services may not justify duplicating each endpoint everywhere, subject to resilience constraints.
Incorrect Answers
B:A gateway endpoint supports private DynamoDB routing without NAT gateway traversal for eligible VPC workloads; it does not grant IAM access by itself; the operational gap would remain because the question requires a way to right-size interface endpoint deployment costs.
C:S3 gateway endpoints route supported S3 traffic privately without NAT processing charges and avoid sending those requests through the NAT gateway path; it is not the action that directly makes the system right-size interface endpoint deployment costs.
D:Local AZ endpoints can reduce cross-AZ dependency for eligible callers while private DNS preserves normal service-hostname resolution; confirm the endpoint configuration and actual DNS resolution path; that still would not establish the requested outcome: right-size interface endpoint deployment costs.
Question 952
A stateful application serves clients across multiple Availability Zones but sends all target traffic to one AZ. During a failure in that AZ, load-balancer capacity is lost despite healthy targets elsewhere. Which reliability adjustment is warranted?
- Spread targets across the required AZs and verify compatible cross-zone forwarding and health configuration
- Add outbound endpoint IP addresses in subnets across at least two supported Availability Zones
- Deploy NAT gateways per active AZ and use same-AZ subnet routes
- Enable private DNS where supported and place endpoint ENIs in the active workload Availability Zones
Explanation
Correct Answer
A:Multi-AZ resilience requires registered usable targets in multiple AZs and load-balancer forwarding behavior that can reach them during a localized failure.
Incorrect Answers
B:Multiple subnet IPs in different AZs reduce dependence on a single AZ for DNS forwarding toward on-premises resolvers; redundant target resolvers should also be reachable; it is not the action that directly makes the system ensure regional load balancing survives an AZ loss.
C:Matching each workload AZ to a local NAT gateway avoids forcing routine egress through another AZ and removes one AZ's gateway as a shared dependency, though gateway hourly costs increase; that still would not establish the requested outcome: ensure regional load balancing survives an AZ loss.
D:Local AZ endpoints can reduce cross-AZ dependency for eligible callers while private DNS preserves normal service-hostname resolution; confirm the endpoint configuration and actual DNS resolution path; the scenario instead needs an action that will ensure regional load balancing survives an AZ loss.
Question 953
A small set of distributed clients sends millions of short-lived TCP connections to one private IP target, and source-port allocation is constrained through a NAT device. Which optimization targets the bottleneck?
- Deploy NAT gateways per active AZ and use same-AZ subnet routes
- Inspect NAT gateway connection and port-allocation metrics for the specific destination tuple
- Reuse application TCP connections or distribute destination tuples and NAT source addresses where feasible
- Clamp TCP MSS to a value compatible with the smallest encapsulated path MTU
Explanation
Correct Answer
C:Reducing churn and spreading simultaneous destination flows can mitigate per-tuple ephemeral-port pressure; simply adding unrelated DNS records does not create more NAT port capacity.
Incorrect Answers
A:Matching each workload AZ to a local NAT gateway avoids forcing routine egress through another AZ and removes one AZ's gateway as a shared dependency, though gateway hourly costs increase; it is not the action that directly makes the system reduce per-destination NAT connection-port pressure.
B:Source port allocation is finite for each NAT gateway address and destination tuple; connection errors under concentrated load can occur before ordinary bandwidth limits are reached; that still would not establish the requested outcome: reduce per-destination NAT connection-port pressure.
D:A successful handshake uses small packets; oversized encrypted data packets may be fragmented or dropped, so an appropriately sized MSS avoids black-hole behavior along the limiting hop; those conditions do not produce the required result: reduce per-destination NAT connection-port pressure.
Question 954
An internal application subnet must deny traffic from one known hostile /24 while allowing a broader corporate /16. Security groups already permit approved corporate traffic. Which subnet-level control can express the explicit deny?
- Permit DNS traffic on the outbound endpoint security group toward the corporate DNS targets
- Configure an ordered network ACL deny on the target subnet with matching destination and protocol scope
- Create the relevant interface VPC endpoints and constrain their endpoint security groups and policies
- Add a numbered DENY entry for the hostile /24 in the relevant network ACL before a broader ALLOW entry
Explanation
Correct Answer
D:Network ACLs evaluate stateless rules in ascending number order and support explicit denies; security groups do not have deny rules. Both ingress and egress paths still need compatible NACL handling.
Incorrect Answers
A:Resolver forwarding generates DNS queries from the endpoint ENIs; restrictive security-group egress and return-path routing can prevent a valid forwarding rule from delivering them; that still would not establish the requested outcome: deny one hostile range within a broader allowlist.
B:A network ACL explicitly rejects matching stateless traffic before later broader allow entries; security groups can only grant permitted flows; the scenario instead needs an action that will deny one hostile range within a broader allowlist.
C:Interface endpoints provide private addresses for supported service APIs, while endpoint security groups and policies can limit which clients and service actions are permitted; those conditions do not produce the required result: deny one hostile range within a broader allowlist.
Question 955
A multitenant ALB service must detect sustained bursts of HTTP requests from a single apparent client IP and automatically block further matching requests. What WAF rule type directly matches the threat?
- Add an AWS WAF rate-based rule with the intended aggregation key and block action
- Use AWS Firewall Manager with an appropriately scoped AWS WAF policy
- Enable AWS WAF request logging and inspect the recorded actions and matching rules
- Apply an AWS WAF managed or custom SQL injection match rule to the application's web ACL
Explanation
Correct Answer
A:Rate-based rules count requests over a configurable evaluation window using the selected aggregation key and apply the configured action after the rate limit is exceeded.
Incorrect Answers
B:Firewall Manager can apply and audit centrally defined security policies across eligible accounts and resources under AWS Organizations, reducing individual distribution policy drift; those conditions do not produce the required result: throttle individual abusive HTTP sources.
C:WAF logs show per-request evaluated rule actions and relevant request attributes, supporting investigation beyond high-level web ACL metric counts; it changes a different aspect of the architecture rather than fulfilling the need to throttle individual abusive HTTP sources.
D:WAF inspects supported HTTP request components and can block matching application-layer attack patterns before the request reaches the origin; the operational gap would remain because the question requires a way to throttle individual abusive HTTP sources.
Question 956
A CloudFront application needs common AWS WAF rules to protect every distribution in an organization. The security account wants centrally enforced policy rather than per-distribution manual changes. Which governance service should orchestrate this?
- Add an AWS WAF rate-based rule with the intended aggregation key and block action
- Enable AWS WAF request logging and inspect the recorded actions and matching rules
- Apply an AWS WAF managed or custom SQL injection match rule to the application's web ACL
- Use AWS Firewall Manager with an appropriately scoped AWS WAF policy
Explanation
Correct Answer
D:Firewall Manager can apply and audit centrally defined security policies across eligible accounts and resources under AWS Organizations, reducing individual distribution policy drift.
Incorrect Answers
A:Rate-based rules count requests over a configurable evaluation window using the selected aggregation key and apply the configured action after the rate limit is exceeded; those conditions do not produce the required result: centrally govern web ACL policies.
B:WAF logs show per-request evaluated rule actions and relevant request attributes, supporting investigation beyond high-level web ACL metric counts; it changes a different aspect of the architecture rather than fulfilling the need to centrally govern web ACL policies.
C:WAF inspects supported HTTP request components and can block matching application-layer attack patterns before the request reaches the origin; the operational gap would remain because the question requires a way to centrally govern web ACL policies.
Question 957
A bank's public HTTPS application needs to reject requests whose query string matches a known SQL injection pattern, while unrelated traffic should continue to reach the application. Which protection layer is most appropriate?
- Apply an AWS WAF managed or custom SQL injection match rule to the application's web ACL
- Add an AWS WAF rate-based rule with the intended aggregation key and block action
- Use AWS Firewall Manager with an appropriately scoped AWS WAF policy
- Enable AWS WAF request logging and inspect the recorded actions and matching rules
Explanation
Correct Answer
A:WAF inspects supported HTTP request components and can block matching application-layer attack patterns before the request reaches the origin.
Incorrect Answers
B:Rate-based rules count requests over a configurable evaluation window using the selected aggregation key and apply the configured action after the rate limit is exceeded; the operational gap would remain because the question requires a way to inspect hostile HTTP query content.
C:Firewall Manager can apply and audit centrally defined security policies across eligible accounts and resources under AWS Organizations, reducing individual distribution policy drift; it is not the action that directly makes the system inspect hostile HTTP query content.
D:WAF logs show per-request evaluated rule actions and relevant request attributes, supporting investigation beyond high-level web ACL metric counts; that still would not establish the requested outcome: inspect hostile HTTP query content.
Question 958
A VPC fleet must prevent engineers from creating an internet gateway attachment in restricted organizational units even if their account IAM permissions otherwise allow it. Which organization-level control can establish that guardrail?
- Run VPC Reachability Analyzer between the source and destination path endpoints
- Search CloudTrail management events for the relevant EC2 gateway-attachment API call
- Attach an AWS Organizations service control policy denying the relevant EC2 internet gateway attachment actions
- Accept the shared Transit Gateway attachment request in the owner account
Explanation
Correct Answer
C:An SCP sets the maximum available permissions for affected principals in member accounts; it does not by itself grant permissions or modify existing route tables.
Incorrect Answers
A:Reachability Analyzer models routing and security configuration to identify a blocking component; it is not a live probe that demonstrates application process health; the operational gap would remain because the question requires a way to restrict internet gateway attachment permissions across accounts.
B:CloudTrail records management-plane API actions such as attachment changes, including caller and timestamp; flow logs cannot identify the actor who modified the control plane; it is not the action that directly makes the system restrict internet gateway attachment permissions across accounts.
D:RAM sharing grants permission to request an attachment, but the TGW owner must accept pending attachment requests unless its attachment auto-acceptance setting handles them; the scenario instead needs an action that will restrict internet gateway attachment permissions across accounts.
Question 959
A highly regulated workload requires access to one S3 bucket through a gateway endpoint, but security requires an additional restriction against using that endpoint for other buckets. Which policy applies at the endpoint?
- Create an S3 gateway endpoint and associate the data subnet route tables
- Create its own supported S3 gateway endpoint and associate the appropriate route tables
- Enable private DNS where supported and place endpoint ENIs in the active workload Availability Zones
- Set the S3 gateway VPC endpoint policy to limit allowed bucket actions and resources
Explanation
Correct Answer
D:Endpoint policies control what AWS service calls can be made through that endpoint; bucket and IAM policies remain independent layers that must also permit the operation.
Incorrect Answers
A:S3 gateway endpoints route supported S3 traffic privately without NAT processing charges and avoid sending those requests through the NAT gateway path; it is not the action that directly makes the system constrain S3 resources reachable through one endpoint.
B:Gateway endpoints are VPC-scoped routing targets and cannot be transited across peering; each participating VPC needs an eligible local endpoint or an alternate supported access pattern; that still would not establish the requested outcome: constrain S3 resources reachable through one endpoint.
C:Local AZ endpoints can reduce cross-AZ dependency for eligible callers while private DNS preserves normal service-hostname resolution; confirm the endpoint configuration and actual DNS resolution path; the scenario instead needs an action that will constrain S3 resources reachable through one endpoint.
Question 960
A private application uses AWS PrivateLink and should accept connection requests only from specific customer AWS accounts. What provider-side configuration controls who can request a connection?
- Expose the service using AWS PrivateLink endpoint services and consumer interface endpoints
- Configure the target group's deregistration delay
- Approve the interface endpoint connection request in the endpoint service owner account
- Restrict endpoint-service allowed principals and require acceptance when appropriate
Explanation
Correct Answer
D:Endpoint-service permissions control which principals can request interface endpoints, while connection acceptance lets the provider approve eligible consumers before traffic flows.
Incorrect Answers
A:PrivateLink offers private service access through endpoints without requiring end-to-end routing between the producer and consumer VPC CIDR ranges; that still would not establish the requested outcome: restrict who may consume a published endpoint service.
B:The deregistration delay sets how long a deregistering target can continue draining existing requests before being fully removed from service; the scenario instead needs an action that will restrict who may consume a published endpoint service.
C:When acceptance is required, a service consumer's endpoint request is not usable until the provider accepts that specific connection; NLB health alone does not approve access; those conditions do not produce the required result: restrict who may consume a published endpoint service.