AWS Secrets Manageris the AWS service designed to store secrets securely and to supportautomatic rotationon a schedule—commonly used for Amazon RDS credentials. Storing credentials in Secrets Manager removes them from source code, enables fine-grained access control, and supports auditability of secret retrieval through CloudTrail. Rotation can be configured to periodically change the database password and update the stored secret automatically, minimizing operational overhead compared to manual rotation processes.

To ensure the credentials are accessibleonly to the application, the correct ECS pattern is to useIAM roles for tasks. A task role can be scoped to allow only secretsmanager:GetSecretValue (and related actions if needed) for the specific secret ARN. Only tasks running with that role can retrieve the secret at runtime, which prevents broad access. This also helps reduce the risk of database administrators sharing plaintext credentials, because the recommended operational model is that humans should not need direct access; the application retrieves the secret programmatically, and access can be limited to break-glass workflows if required.

Systems Manager Parameter Store can store encrypted parameters, but Secrets Manager provides stronger native secret lifecycle features (notably rotation) for databases. Inline policies (Option B) are not necessary; managed or attached policies on the task role achieve the same goal with cleaner administration.

The correct solution is option D because it effectively prevents direct access to the internet-facing ALB while allowing legitimate traffic that originates from Amazon CloudFront. By configuring CloudFront to include a custom HTTP header (such as X-Shared-Secret) in all origin requests, and then configuring ALB listener rules to only forward requests that contain the expected header value, the ALB will reject any requests that bypass CloudFront.

This approach is a documented AWS best practice when CloudFront is placed in front of an ALB and AWS WAF is associated with the CloudFront distribution. AWS WAF only evaluates traffic that flows through CloudFront; therefore, preventing direct access to the ALB is critical to ensure that all requests are inspected by the web ACL.

Option A is invalid because CloudFront does not support AWS PrivateLink endpoints as origins. Option B is incorrect because CloudFront cannot use an internal ALB as an origin; CloudFront requires a publicly reachable origin. Option C is not recommended because CloudFront IP ranges change frequently, making IP-based allow lists operationally complex and error-prone, and AWS does not provide a supported CloudFront prefix list for ALB listener rules.

AWS Security Specialty guidance explicitly recommends using custom origin headers to restrict ALB access to CloudFront-only traffic, making option D the correct and secure solution.

The most efficient approach is to useAWS Configbecause Config is designed for continuous compliance evaluation and can automatically triggermanaged remediationwhen a resource drifts from the desired state. A managed Config rule that detects when CloudTrail is not logging, combined with theAWS-EnableCloudTrailremediation action, provides an automated way to re-enable CloudTrail without building and maintaining custom event processing code. This is especially valuable in multi-Region environments because Config can evaluate configurations across Regions and enforce the intended posture consistently.

Option B is illogical: triggering on StartLogging does not help when CloudTrail is turned off. Option C is not as operationally efficient because CloudWatch alarms are not the standard mechanism for reacting to CloudTrail API events; EventBridge is the proper event bus for API call events, but you would still be writing and maintaining Lambda logic and multi-Region plumbing. Option D is manual and delayed, not automated remediation.

Therefore, AWS Config with a managed rule and the AWS-provided remediation to enable CloudTrail is the most maintainable and efficient solution.

Amazon S3 Block Public Access provides centralized controls to prevent public access through bucket policies and ACLs. AWS Certified Security – Specialty guidance recommends enabling Block Public Access to reduce accidental exposure and to enforce guardrails that override public grants. Enabling Block Public Access on the bucket removes current public exposure when combined with correcting policies/ACLs and prevents future misconfiguration. To ensure the bucket cannot be made public again, the security engineer must prevent principals from disabling Block Public Access. An SCP that denies s3:PutPublicAccessBlock prevents changes that would remove or weaken the PublicAccessBlock configuration, enforcing the guardrail across the OU or account. Options A and D do not directly address public exposure control. Option B denies object reads but does not ensure public access cannot be re-enabled; it also does not address the root misconfiguration pathways and could disrupt legitimate access patterns. Option C specifically combines the correct preventive control (PublicAccessBlock) with organizational enforcement to stop future reversal.

Referenced AWS Specialty Documents:

AWS Certified Security – Specialty Official Study Guide

Amazon S3 Block Public Access

AWS Organizations SCP Guardrails for S3 Controls

The policy currently grants s3:GetObject but targets thebucket ARN(arn:aws:s3:::DOC-EXAMPLE-BUCKET). For Amazon S3, object-level actions such asGetObjectmust referenceobject ARNs, not the bucket ARN. The correct resource pattern is the bucket ARNwith /*appended (for example, arn:aws:s3:::DOC-EXAMPLE-BUCKET/*) so the permission applies to objects within the bucket. Without this, S3 evaluates the request against a resource that does not match the requested object, resulting in an access denial even though the action appears correct.

The other options do not address the root cause. Expanding actions (Option B) is unnecessary and overly permissive, and it still would not fix the incorrect resource ARN for object reads. Changing principals or removing conditions (Option A) is not required just to allow reads—Lambda typically accesses S3 using the function’sexecution role, and bucket policies are commonly used for cross-account or service-based access control, but the immediate failure here is the mismatch between s3:GetObject and the bucket-only resource. Option D is invalid because it inverts principal/service usage and sets an incorrect resource type for S3 authorization.

Option C best meets the requirements because Amazon GuardDuty provides Kubernetes-focused threat detection for Amazon EKS by analyzingEKS control plane audit logs(EKS Protection) and combining that signal withruntime telemetryfrom the worker nodes (Runtime Monitoring). EKS audit logs capture Kubernetes API activity and authorization decisions, allowing GuardDuty to detect suspicious cluster actions such as unusual API calls, unexpected access patterns, or indicators of compromise within the cluster. Runtime Monitoring extends coverage tooperating system/process activity, network connections, and file activityon the nodes, which directly aligns with the need to monitor OS, networking, and file events in addition to audit logs.

For notifications, GuardDuty generatesfindingsthat can be delivered throughAmazon EventBridgerules. EventBridge can route relevant GuardDuty findings to anAmazon SNS topic, and SNS can sendemail alertsto the security team by subscribing the team’s mailing list to the topic. This approach is fully managed, near real time, and avoids building custom log-parsing pipelines while still providing actionable alerts based on GuardDuty’s curated EKS threat detections.

For encrypted RDS snapshots that must be shared across accounts and still bedecryptablein the target accounts, you should use acustomer managed KMS keyand explicitly grant cross-account use of that key. AWS-managed default keys (Option A/C) generally cannot be shared for cross-account decryption in the same flexible way as customer managed keys, and you cannot “copy” an AWS-managed key to another account. Likewise, you cannot “import” an existing KMS key into another account via Lambda as described in Option B; KMS keys are account-scoped resources and are not copied between accounts like that.

With acustomer managed key (CMK), the key policy (and/or grants) can allow principals in the DR accounts to use the key for the required cryptographic operations (for example, kms:Decrypt, kms:CreateGrant, and relevant describe permissions). Then, when the snapshot is shared and copied/used in the destination account during a DR event, the destination account can decrypt it because it has been granted permission to use the same CMK. This approach is the standard AWS pattern for cross-account encrypted snapshot sharing and meets both encryption and recoverability requirements with strong governance and auditability through CloudTrail.

Amazon CloudFront includes a built-in geo restriction feature that allows content to be allowed or denied based on the viewer’s country. According to AWS Certified Security – Specialty documentation, CloudFront geo restriction is the most cost-effective method for country-based blocking because it does not require AWS WAF or additional rule processing.

AWS WAF geo match rules incur additional cost and are more appropriate when advanced inspection or layered security controls are required. IP-based blocking is impractical due to frequent IP changes. Geolocation headers do not enforce access control.

CloudFront geo restriction is evaluated at the edge and efficiently blocks disallowed countries with minimal latency and cost.

Referenced AWS Specialty Documents:

AWS Certified Security – Specialty Official Study Guide

Amazon CloudFront Geo Restriction

AWS Edge Security Best Practices

EC2 Instance Connect can performserver/host authenticity checksby validating the instance’s SSHhost keyagainst atrusted host keyssource. When you rotate the instance’s host keys, the host presents anewfingerprint. If the trusted host keys source still contains theoldhost key, connections that enforce host key verification will fail with ahost key validationerror. The fix is to update the trusted host key record so the new host key fingerprint is recognized as valid. Therefore, the correct action is toupload the new host keyto the trusted host keys database used for EC2 Instance Connect host key verification.

Option A is unrelated: AWS KMS does not store or manage SSH host keys for EC2 Instance Connect validation. Option C is for AWS Systems Manager managed instances and has no effect on SSH host key validation. Option D rotatesuser/client authentication keys(SSH key pair used to log in) but does not resolve a failure that occurs specifically because theserver host keychanged and is no longer trusted. Updating the trusted host keys database restores the expected trust chain and allows Instance Connect to work again with the rotated host keys.

The company needs centralized, scalable protection acrossmany accountsfor bothexisting and futureAPI Gateway resources, with minimal ongoing effort.AWS Firewall Manageris specifically designed for this: it can centrally deploy and enforceAWS WAFprotections across AWS Organizations. By creating a Firewall ManagerWAF policy, the security team defines a single set of controls (for example, AWS Managed Rules for SQL injection protection andAWS Bot Control) and applies them automatically to in-scope resources across member accounts.

Critically, Firewall Manager can be configured toauto-remediate noncompliant resources, ensuring that if new API Gateway stages are created later, they are automatically brought under the policy without manual per-account work. This directly meets the “existing and future resources” requirement.

Options A, C, and D introduce higher operational overhead: per-API ACL creation plus AWS Config remediation (A) is more moving parts; Service Catalog plus detection/remediation (C) is indirect and heavy; and Security Hub + EventBridge + Lambda automation (D) is custom engineering and maintenance. Firewall Manager is the AWS-native centralized governance solution for multi-account WAF rollout and enforcement.