SOA-C03 AWS Certified CloudOps Engineer - Associate Questions and Answers

To analyze and visualize custom statistics such as the p90 latency (90th percentile), a CloudWatch metric must be generated from the log data. The correct method is to create a metric filter that extracts the latency value from each log event and publishes it as a CloudWatch metric. Once the metric is published, percentile statistics (p90, p95, etc.) can be displayed in CloudWatch dashboards or alarms.

AWS documentation states:

“You can use metric filters to extract numerical fields from log events and publish them as metrics in CloudWatch. CloudWatch supports percentile statistics such as p90 and p95 for these metrics.”

Contributor Insights (Option A) is for analyzing frequent contributors, not numeric distributions. Subscription filters (Option C) are used for log streaming, and Application Insights (Option D) provides monitoring of application health but not custom p90 statistics. Hence, Option B is the CloudOps-aligned, minimal-overhead solution for percentile latency monitoring.

High availability requires removing single-AZ risk and eliminating manual recovery. The AWS Reliability best practices state to design for multi-AZ and automatic healing: Auto Scaling “helps maintain application availability and allows you to automatically add or remove EC2 instances” (AWS Auto Scaling User Guide). The Reliability Pillar recommends to “distribute workloads across multiple Availability Zones” and to “automate recovery from failure” (AWS Well-Architected Framework – Reliability Pillar). Attaching the Auto Scaling group to an ALB target group enables health-based replacement: instances failing load balancer health checks are replaced and traffic is routed only to healthy targets. Using an AMI in a launch template ensures consistent, repeatable instance configuration (AWS EC2 Launch Templates). Options A and C keep all instances in a single Availability Zone and rely on manual or ad-hoc restarts, which do not meet high-availability or resiliency goals. Option B only scales vertically and adds a restart rule; it neither removes the single-AZ failure domain nor provides automated replacement. Therefore, creating a multi-AZ EC2 Auto Scaling group with a launch template and attaching it to the ALB target group (Option D) is the CloudOps-aligned solution for resilience and business continuity.

Comprehensive and Detailed Explanation From Exact Extract of AWS CloudOps Doocuments:

For high availability and fast failover, ElastiCache for Redis supports replication groups with Multi-AZ and automatic failover. CloudOps guidance states that a primary node can be paired with one or more replicas across multiple Availability Zones; if the primary fails, Redis automatically promotes a replica to primary in seconds, thereby minimizing RTO. This architecture maintains in-memory data continuity without waiting for backup restore operations. Backups (Options B and D) provide durability but require restore and re-warm procedures that increase RTO and may impact application latency. Switching engines (Option A) to Memcached does not provide Redis replication/failover semantics and would not inherently improve resilience for this use case. Therefore, creating a read replica in a different AZ and enabling Multi-AZ with automatic failover is the prescribed CloudOps pattern to increase resilience and achieve the lowest practical RTO for Redis caches.

The AWS Cloud Operations and Governance documentation defines that enabling Organizational View in AWS Health allows the management account in AWS Organizations to view and aggregate health events from all member accounts.

This feature provides a single-pane-of-glass view of service health issues, account-specific events, and planned maintenance across the organization — without requiring additional automation or data pipelines.

Alternative options (B, C, and D) require custom integration and ongoing maintenance. CloudTrail does not natively forward AWS Health events, and custom Lambda or DynamoDB approaches increase complexity.

Therefore, Option A — enabling the Organizational View feature in AWS Health — is the most operationally efficient and AWS-recommended solution.

CloudWatch Logs data protection provides native redaction/masking of sensitive data at ingestion and query. AWS documentation states it can “detect and protect sensitive data in logs” using data identifiers, and that authorized users can “use the unmask action to view the original data.” Creating a data protection policy on the log group masks PII by default for all viewers, satisfying the requirement to redact personal information. Granting only the security team permission to invoke the unmask API operation ensures that unredacted content is restricted. Option B (KMS) encrypts at rest but does not redact fields; encryption alone does not prevent plaintext visibility to authorized readers. Options A and D add complexity and latency, move data out of CloudWatch, and do not provide default inline redaction/unmask controls in CloudWatch itself. Therefore, the CloudOps-aligned, managed solution is to use CloudWatch Logs data protection with appropriate data identifiers and unmask permissions limited to the security team.

According to the AWS Cloud Operations and Disaster Recovery documentation, the active-active multi-Region architecture provides the lowest possible RTO and RPO among all disaster recovery strategies. In this approach, workloads are deployed and actively running in multiple AWS Regions simultaneously. All data is continuously replicated in real time between Regions using fully managed replication services, ensuring zero data loss (zero RPO).

Because both Regions are active and capable of handling requests, failover between them is instantaneous, meeting the RTO of less than 10 minutes. Amazon Route 53 is used with weighted or latency-based routing policies and health checks to automatically route traffic away from an impaired Region to the healthy Region without manual intervention.

In contrast:

Pilot Light Architecture maintains only a minimal copy of the environment in the secondary Region. It requires time to scale up infrastructure during a disaster, resulting in longer RTO and potential data loss (non-zero RPO).

Warm Standby Architecture keeps partially running infrastructure in the secondary Region. Although faster than pilot light, it still requires scaling and synchronization, resulting in higher RTO and RPO compared to active-active.

Backup and Restore (option D) relies on periodic backups and restores data when needed. This approach has the highest RTO and RPO, unsuitable for mission-critical workloads demanding high availability and zero data loss.

Therefore, based on AWS-recommended disaster recovery strategies outlined in the AWS Cloud Operations and Disaster Recovery Guide, the Active-Active Multi-Region architecture (Option C) is the only approach that guarantees RTO <10 minutes and RPO = 0, achieving continuous availability and business continuity across Regions.

Comprehensive and Detailed Explanation From Exact Extract of AWS CloudOps Doocuments:

With simple scaling policies, an Auto Scaling group performs one scaling activity when the alarm condition is met, then observes a default cooldown period (300 seconds) before another scaling activity of the same type can begin. CloudOps guidance explains that cooldown prevents rapid successive scale-outs by allowing time for the newly launched instance(s) to register with the load balancer and impact the metric. Even if the alarm breaches multiple times during the cooldown window, the group waits until the cooldown completes before evaluating and acting again. In this case, although RequestCountPerTarget exceeded the threshold twice within 180 seconds, the group will launch a single instance and then wait for cooldown before any additional scale-out can occur. Options A, C, and D do not reflect the behavior of simple scaling with cooldowns; A describes step/target-tracking-like behavior, and C/D are not Auto Scaling mechanics.

According to the AWS Cloud Operations and Database Reliability documentation, Amazon RDS Multi-AZ deployments provide high availability and automatic failover by maintaining a synchronous standby replica in a different Availability Zone.

In the event of instance failure, planned maintenance, or Availability Zone outage, Amazon RDS automatically promotes the standby to primary with minimal downtime (typically less than 60 seconds). The failover is transparent to applications because the DB endpoint remains the same.

By contrast, read replicas (Option B) are asynchronous and do not provide automated failover. Auto Scaling (Option C) applies to EC2, not RDS. RDS Proxy (Option D) improves connection management but does not add redundancy.

Thus, Option A — converting the RDS instance into a Multi-AZ deployment — delivers the required high availability and business continuity with minimal operational effort.

Comprehensive and Detailed Explanation From Exact Extract of AWS CloudOps Doocuments:

VPC Flow Logs show the request arriving and being ACCEPTed on dstport 8080 and the corresponding response being REJECTed on the return path to the client’s ephemeral port (59003). AWS networking guidance states that security groups are stateful (return traffic is automatically allowed) while network ACLs are stateless and require explicit inbound and outbound rules for both directions. CloudOps operational guidance for VPC networking further notes that when you allow an inbound request (for example, TCP 8080) through a subnet’s network ACL, you must also allow the outbound ephemeral port range (typically 1024–65535) for the response traffic; otherwise, the return packets are dropped and appear as REJECT in flow logs. The observed pattern—request accepted to 8080, response rejected to 59003—matches a missing outbound ephemeral-range allow on the subnet’s NACL. Therefore, the cause is the subnet NACL, not security groups or on-premises ACLs. The remediation is to add an outbound ALLOW rule on the NACL for the appropriate ephemeral TCP port range back to the on-premises CIDR (and the corresponding inbound rule if asymmetric).

According to the AWS Cloud Operations and Identity Management documentation, when configuring federation between IAM Identity Center (formerly AWS SSO) and an external SAML 2.0 identity provider, two key prerequisites are required:

The IAM Identity Center SAML metadata file — This is uploaded to the external IdP to establish trust, define SAML endpoints, and enable identity federation.

The IdP metadata (including the public X.509 certificate) — This information is imported into IAM Identity Center to validate authentication assertions and encryption signatures.

IAM Identity Center and the IdP exchange this metadata to mutually establish secure, bidirectional federation.

Network-level details such as IP addresses (Option C) are unnecessary. Root access (Option D) or permissions to member accounts (Option E) are not required; only Control Tower or IAM administrative permissions in the management account are needed for setup.

Thus, the correct answer is A and B — the SAML metadata from both sides is required for federation.

As per the AWS Cloud Operations and Event Monitoring documentation, the most efficient method for event-driven notification is to use Amazon EventBridge to detect specific EC2 API events and trigger a Simple Notification Service (SNS) alert.

EventBridge continuously monitors AWS service events, including RunInstances, which signals the creation of new EC2 instances. When such an event occurs, EventBridge sends it to an SNS topic, which then immediately emails subscribed recipients — in this case, the architecture team.

This combination provides real-time, serverless notifications with minimal management. SQS (Option C) is designed for queue-based processing, not direct user alerts. User data scripts (Option A) and custom polling with Lambda (Option D) introduce unnecessary operational complexity and latency.

Hence, Option B is the correct and AWS-recommended CloudOps design for immediate launch notifications.

Comprehensive and Detailed Explanation From Exact Extract of AWS CloudOps Doocuments:

Amazon Aurora supports up to 15 Aurora Replicas that share the same storage volume and provide read scaling and improved availability. Official guidance states that replicas “offload read traffic from the writer” and that you should direct read-only workloads to the reader endpoint, reducing CPU pressure and connection counts on the primary. Aurora also supports Replica Auto Scaling through Application Auto Scaling policies using metrics such as CPU utilization or connections to add or remove replicas automatically. This design addresses both high CPU and maximum connections by moving reporting traffic to read replicas while keeping a single write primary for OLTP. Option B creates a separate cluster with independent storage, increasing operational overhead and data synchronization complexity. Options C and D introduce application-layer caching changes that may not guarantee data freshness or relieve the write node directly. Therefore, adding read replicas and routing reporting to the reader endpoint, with auto scaling based on load, is the least intrusive, CloudOps-aligned way to stabilize performance.

According to the AWS Cloud Operations and ElastiCache documentation, cache evictions occur when the cache runs out of memory and must remove items to make space for new data.

To reduce evictions and retain frequently accessed items, AWS recommends increasing the total available memory — either by scaling up to larger node types or scaling out by adding shards/nodes. Migrating to a cluster with larger nodes is the simplest and most efficient solution because it immediately expands capacity without architectural changes.

Adjusting TTL (Options B and C) controls expiration timing, not memory allocation. Adding a single node (Option A) may help, but redistributing data requires resharding, introducing more complexity.

Thus, Option D provides the lowest operational overhead and ensures high cache hit rates by increasing total cache memory.

Per the AWS Cloud Operations and Event Management documentation, Amazon EventBridge provides native scheduling capabilities that can trigger events at defined intervals—such as weekly, daily, or cron-based schedules.

Creating an EventBridge rule that runs every Saturday and publishes a message to an SNS topic fully automates the notification process without maintaining servers or manual jobs. This approach is serverless, highly reliable, and fully managed by AWS.

By contrast:

EC2 cron jobs (Option A) require instance management, patching, and cost overhead.

SNS subscriptions (Option C) handle message delivery, not scheduling.

Step Functions (Option D) are designed for complex workflows, not simple scheduled triggers.

Thus, Option B provides the most operationally efficient CloudOps solution by integrating EventBridge scheduled events with SNS topics for automated, recurring notifications.

Comprehensive and Detailed Explanation From Exact Extract of AWS CloudOps Documents:

AWS Lambda automatically publishes operational metrics to Amazon CloudWatch, including Duration, which represents the time a function invocation takes to run. To alert when processing time exceeds 10 seconds, the most direct and operationally efficient solution is to create a CloudWatch alarm on the Lambda Duration metric and configure the alarm action to publish to an SNS topic. This meets the monitoring requirement without requiring log parsing or additional query mechanisms.

Option B fits best because it uses the built-in metric stream for Lambda observability: CloudWatch metrics are near real-time, require minimal configuration, and alarms are a standard CloudOps practice for proactive notification. The alarm can be configured with the appropriate statistic (for example, Maximum or p99 via metric math where applicable) and a threshold of 10,000 milliseconds, ensuring the operations team is notified before performance degrades further.

Option A is incorrect because PostRuntimeExtensionsDuration is not the primary runtime metric for function execution time, and extracting runtime from logs is unnecessary. Option C changes the function timeout to 10 seconds, which would cause failures rather than simply notifying on slow executions. Option D is more operationally complex because it relies on log queries; CloudWatch alarms are more straightforward when a native metric exists.

As per the AWS Cloud Operations and Content Delivery documentation, CloudFront determines cache behavior by evaluating both origin headers (e.g., Cache-Control and Expires) and distribution-level TTL settings.

When Cache-Control max-age conflicts with the Maximum TTL configured in CloudFront, the shorter TTL value takes precedence. This results in CloudFront caching content for only 5 minutes instead of 1 hour, despite the origin headers suggesting a longer duration.

AWS documentation explicitly states: “When both origin cache headers and CloudFront TTL settings are defined, CloudFront uses the most restrictive caching period.” This mismatch causes the perceived performance drop, as CloudFront frequently revalidates content.

Therefore, Option D is correct — cache-duration settings conflict with each other, leading to unexpected caching behavior.

The AWS Cloud Operations and Security documentation specifies that for Amazon CloudFront, automatic certificate renewal is only supported for certificates issued by AWS Certificate Manager (ACM). When a certificate is managed by ACM and validated through DNS validation, ACM automatically renews the certificate before expiration without requiring manual intervention.

Option A ensures that the certificate is issued and managed by ACM, enabling full integration with CloudFront. Option E (DNS validation) is essential for automation; AWS performs revalidation automatically as long as the DNS validation record remains in place.

By contrast, email validation (Option D) requires manual user confirmation upon renewal, which prevents automatic renewals. Certificates issued by third-party certificate authorities (Option B) are manually managed and must be reimported into ACM after renewal. CloudFront does not have a direct feature (Option C) to renew certificates; it relies on ACM’s lifecycle management.

Thus, combining ACM-issued certificates (A) with DNS validation (E) ensures continuous, automated renewal with no downtime or human action required.

Comprehensive and Detailed Explanation From Exact Extract of AWS CloudOps Documents:

The correct answer is B because cross-account role assumption requires two explicit permissions. AWS CloudOps documentation states that the target role must trust the principal, and the principal must be allowed to call sts:AssumeRole.

In the member account, the role’s trust policy must list the IAM group ARN (or the identity account) as a trusted principal. In the identity account, the IAM group must have an inline or attached policy that allows the sts:AssumeRole action for the target role ARN. This dual configuration enables secure and controlled cross-account access.

Option A is incorrect because trust policies cannot be attached to IAM groups. Option C is incorrect because sts:PassRole is used for passing roles to AWS services, not for assuming roles. Option D is incorrect because roles do not grant permissions via inline policies to principals.

This approach aligns precisely with AWS CloudOps guidance for multi-account IAM design.

The AWS Cloud Operations and Infrastructure-as-Code documentation specifies that when using AWS CloudFormation, resources are created in parallel by default unless explicitly ordered using DependsOn.

If a custom resource (Lambda) depends on another resource (like an EC2 instance) to exist before execution, a DependsOn attribute must be added to enforce creation order. This ensures the EC2 instance is launched and available before the custom resource executes its automation logic.

Updating the service token (Option B) doesn’t affect order of execution. The cfn-response module (Option C) handles callback communication but not sequencing. Fn::If (Option D) is for conditional creation, not dependency control.

Therefore, Option A is correct — adding a DependsOn attribute guarantees that CloudFormation provisions the EC2 instance before executing the Lambda custom resource.

According to CloudOps automation and monitoring best practices, CloudWatch dashboards should be provisioned as infrastructure-as-code (IaC) resources using AWS CloudFormation to ensure consistency, repeatability, and version control. AWS CloudFormation supports the AWS::CloudWatch::Dashboard resource, where the DashboardBody property accepts a JSON object describing widgets, metrics, and layout.

By exporting the existing dashboard configuration as JSON and embedding it into the CloudFormation template, every deployment of the application automatically creates its corresponding dashboard. This method aligns with the CloudOps requirement for automated deployment and operational visibility within the same stack lifecycle.

AWS documentation explicitly states:

“Use the AWS::CloudWatch::Dashboard resource to create a dashboard from your template. You can include the same JSON you use to define a dashboard in the console.”

Option A requires manual execution. Options C and D incorrectly reference or reuse existing dashboards, failing to produce unique, deployment-specific dashboards.

Comprehensive and Detailed Explanation From Exact Extract of AWS CloudOps Documents:

The correct answer is D because target tracking scaling policies are designed to automatically maintain a specific metric at a desired target value with the least administrative effort. AWS CloudOps documentation states that target tracking works like a thermostat, continuously adjusting capacity to keep the selected metric close to the defined target.

By specifying a target value of 40% average CPU utilization, the Auto Scaling group automatically scales out or in based on real-time demand without requiring manual thresholds, multiple alarms, or predefined scaling steps. This approach is fully managed by AWS and is the most operationally efficient option.

Option A is incorrect because scheduled scaling is based on predictable traffic patterns and cannot respond dynamically to real-time load changes. Option B is incorrect because simple scaling reacts only after an alarm is triggered and does not continuously maintain a target value. Option C is less efficient because step scaling requires additional configuration and tuning of scaling steps.

AWS CloudOps best practices clearly recommend target tracking scaling policies for maintaining steady performance metrics with minimal operational overhead.

Comprehensive and Detailed Explanation From Exact Extract of AWS CloudOps Documents:

The correct answer is C because selective CloudFront invalidation ensures that only updated JavaScript files are refreshed across edge locations. AWS CloudOps documentation explains that invalidations remove cached objects so that CloudFront fetches the latest version from the origin on the next request.

Invalidating only changed files minimizes cost, reduces operational impact, and avoids unnecessary origin requests. This approach ensures users always receive the latest application assets without degrading backend performance.

Option A is incorrect because setting TTLs to 0 forces CloudFront to query the origin for every request, increasing load on the ALB and EC2 instances. Option B is inefficient and costly because invalidating all files is unnecessary. Option D removes the benefits of CloudFront caching and increases latency.

AWS CloudOps best practices recommend targeted invalidations during deployments to balance performance, cost, and correctness.

According to the AWS Cloud Operations and Security documentation, AWS CloudTrail is the authoritative service for recording API activity across all AWS services within an account.

When an access key is compromised, CloudTrail logs all API requests made using that key, including details such as:

The user identity (access key ID) that made the request,

The service, operation, resource, and timestamp affected, and

The source IP address and region of the request.

By searching the CloudTrail event history for the specific access key ID, the CloudOps engineer can identify every action performed by that key during the suspected breach window.

Other options are incorrect:

EventBridge (A) is event-driven, not historical.

CloudWatch Logs (B) monitors system logs, not AWS API activity.

VPC Flow Logs (D) track network-level traffic, not API calls.

Therefore, the correct solution is Option C — using AWS CloudTrail event history to audit and trace all actions executed via the compromised access key.

Comprehensive and Detailed Explanation From Exact Extract of AWS CloudOps Documents:

The correct answer is A because AWS Systems Manager Patch Manager natively supports tag-based targeting, which automatically includes both existing and future instances that match specified tag criteria. AWS CloudOps documentation states that patch policies can target managed nodes by instance tags, allowing administrators to dynamically scope patching operations without additional automation.

By defining the patch policy target as instances with an environment tag value of “Prod,” Patch Manager automatically applies patch baselines to all matching instances. Any new EC2 instance launched with the same tag is included automatically, requiring no manual intervention or additional services. This approach delivers the least operational overhead while remaining fully scalable and compliant.

Options B, C, and D are incorrect because they introduce unnecessary complexity by adding AWS Lambda functions, resource groups, or EventBridge rules. AWS CloudOps best practices emphasize using native Systems Manager capabilities whenever possible to reduce operational burden and failure points.