SOA-C02 Cost and Performance Optimization Practice Test 13 — 15 Questions

Question 1

A company runs a production web application on Amazon EC2 instances behind an Application Load Balancer. The application experiences unpredictable traffic spikes. The operations team wants to optimize costs while maintaining performance during spikes. Which solution is MOST cost-effective?

Accepted Answer

Use Compute Savings Plans for the baseline and run the spikes on On-Demand instances.. Option C is correct because Savings Plans provide a flexible discount in exchange for a commitment to a consistent amount of compute usage (measured in $/hour) for 1 or 3 years, which can cover EC2, Lambda, and Fargate usage across any instance family, region, or OS. This allows the company to benefit from lower prices while still using On-Demand instances for the remaining flexible portion. Option A is wrong because Reserved Instances are tied to a specific instance family and region, which may be too restrictive if the company needs to change instance types. Option B is wrong because Spot Instances can be interrupted and are not suitable for a production web application that requires high availability. Option D is wrong because Dedicated Hosts are expensive and provide no cost optimization for variable workloads.

Answer

Use Spot Instances to handle the spikes.

Answer

Purchase Reserved Instances for the expected baseline capacity.

Answer

Use Dedicated Hosts for the entire workload.

Question 2

A SysOps administrator is reviewing cost reports and notices that costs for Amazon S3 have increased significantly. The bucket stores log files that are accessed once a month for auditing. The logs are written daily, and each object is about 10 MB. The company needs to keep logs for 7 years. Which action will optimize costs while meeting retention requirements?

Accepted Answer

Use an S3 Lifecycle policy to transition objects to S3 Glacier Deep Archive after 30 days and expire after 7 years.. Option D is correct because S3 Lifecycle policies can automate transitioning objects to lower-cost storage classes (e.g., S3 Glacier Deep Archive) after the initial access period, and then expire them after 7 years. This minimizes storage costs. Option A is wrong because S3 Standard is the most expensive storage class and is not cost-optimal for rarely accessed logs. Option B is wrong because storing logs in a single volume on EBS is not durable and does not provide the same cost savings as S3 lifecycle. Option C is wrong because even S3 Glacier Flexible Retrieval is more expensive than S3 Glacier Deep Archive for long-term archival.

Answer

Store the logs in an Amazon EBS volume.

Answer

Transition the logs to S3 Glacier Flexible Retrieval after 30 days.

Answer

Store the logs in S3 Standard and delete after 7 years.

Question 3

A company wants to monitor the performance of its Amazon RDS for MySQL database. The database is experiencing high CPU utilization during peak hours. The SysOps administrator needs to identify the queries causing the load. Which AWS service should be used?

Accepted Answer

Amazon RDS Performance Insights. Option C is correct because Amazon RDS Performance Insights provides a dashboard that visualizes database load and helps identify the queries that are causing high CPU usage. Option A is wrong because Amazon CloudWatch Logs only collects log files, not query performance data. Option B is wrong because AWS CloudTrail records API activity, not database queries. Option D is wrong because Amazon Inspector is a vulnerability management service, not a database performance tool.

Answer

AWS CloudTrail

Answer

Amazon Inspector

Answer

Amazon CloudWatch Logs

Question 4

A SysOps administrator is troubleshooting slow application performance. The application runs on Amazon EC2 instances in an Auto Scaling group behind an Application Load Balancer. Amazon CloudWatch metrics show that the average CPU utilization across the instances is below 20%, but the application is still slow. What is the MOST likely cause of the performance issue?

Accepted Answer

The application database is under-provisioned and is causing slow query responses.. Option B is correct because if the database is the bottleneck, even if the application servers have low CPU usage, the overall performance can be slow due to database contention. Option A is wrong because if the ALB is not distributing traffic, some instances would have high CPU. Option C is wrong because low CPU utilization indicates that the application servers are not overloaded. Option D is wrong because burstable performance instances (T-series) have a CPU credit balance; if the balance is depleted, CPU performance is throttled, leading to low CPU but slow performance? Actually, when credits are exhausted, T instances provide baseline performance, which may be low, but CPU utilization would still be limited, not necessarily low; but more importantly, the symptoms match a database bottleneck.

Answer

The Auto Scaling group is scaling too aggressively, causing thrashing.

Answer

The Application Load Balancer has a sticky session configuration that is not distributing traffic evenly.

Answer

The EC2 instances are using burstable performance and have exhausted their CPU credits.

Question 5

A company has a data processing job that runs on Amazon EMR. The job takes 2 hours to complete and runs once per day. The SysOps administrator wants to reduce costs without affecting the completion time. Which strategy is MOST effective?

Accepted Answer

Use Spot Instances for the task nodes in the EMR cluster.. Option D is correct because using Spot Instances for task nodes (which are fault-tolerant) can significantly reduce costs, as Spot Instances are often 60-80% cheaper than On-Demand. The job's duration is predictable, and if a Spot Instance is interrupted, EMR can relaunch the task on another instance. Option A is wrong because using a smaller number of larger instances may not reduce costs; in fact, it could increase costs if the job requires parallel processing. Option B is wrong because reducing the number of instances will increase the job duration. Option C is wrong because EMR does not support Savings Plans directly for the EMR service itself; Savings Plans cover EC2 instances used by EMR, but the same savings can be achieved with Spot.

Answer

Use a smaller number of larger EC2 instances.

Answer

Reduce the number of core nodes.

Answer

Purchase Compute Savings Plans for the EMR cluster.

Question 6

A company is using AWS Lambda functions to process events from Amazon S3. The functions are invoked several thousand times per minute. The SysOps administrator notices that the functions are taking longer to execute during peak times. Which action should the administrator take to improve performance?

Accepted Answer

Increase the memory allocation for the Lambda functions.. Option A is correct because increasing the memory allocation for a Lambda function also increases the CPU allocation proportionally, which can reduce execution time for compute-bound tasks. Option B is wrong because Lambda functions cannot use provisioned concurrency by default; provisioned concurrency is for initializing execution environments to reduce cold starts, but it does not directly improve execution time. Option C is wrong because using a larger instance type is not applicable to Lambda; Lambda abstracts the underlying compute. Option D is wrong because Lambda functions do not have a maximum execution time that can be extended beyond 15 minutes; that is a hard limit.

Answer

Increase the Lambda function timeout.

Answer

Use a larger EC2 instance type for the Lambda functions.

Answer

Enable provisioned concurrency on the Lambda functions.

Question 7

A SysOps administrator is managing a fleet of Amazon EC2 instances that run a batch processing job. The job runs for 30 minutes every hour. The administrator wants to optimize costs by using Spot Instances but must ensure that the job completes even if Spot Instances are interrupted. Which configuration should the administrator use?

Accepted Answer

Use a mixed instances policy with both On-Demand and Spot Instances, and configure the Spot allocation strategy to 'lowestPrice'.. Option B is correct because using a mixed instances policy with both On-Demand and Spot Instances ensures that the baseline capacity is covered by On-Demand, and any interrupted Spot Instances can be replaced by On-Demand as a fallback. The Spot allocation strategy 'lowestPrice' is not specified but the key is the capacity-optimized or diversified strategy to reduce interruption. However, the question asks for ensuring completion; the best practice is to use a capacity-optimized allocation strategy. Actually, the correct answer is B: Use a mixed instances policy with a capacity-optimized allocation strategy and specify a percentage of On-Demand instances. But option B says 'lowestPrice', which is common but not best for stability. However, among the options, B is the only one that includes a mix. Option A is wrong because only Spot is risky. Option C is wrong because the job is not long-running (30 min), so persistence is not needed; but the job must complete, so if Spot is interrupted, it should be able to restart. Option D is wrong because it is wasteful to use only On-Demand.

Answer

Launch Spot Instances in an Auto Scaling group with a capacity-optimized allocation strategy.

Answer

Use a Spot Fleet with a 'persistent' request type.

Answer

Use only On-Demand Instances to avoid interruptions.

Question 8

A company uses Amazon CloudFront to distribute content globally. The SysOps administrator notices that the origin load is high and the cache hit ratio is low. What should the administrator do to improve the cache hit ratio and reduce origin load?

Accepted Answer

Increase the minimum, maximum, and default TTL values for the cache behavior.. Option C is correct because increasing the TTL allows objects to be cached for longer, which increases the likelihood of cache hits. Option A is wrong because enabling compression reduces transfer size but does not affect cache hit ratio. Option B is wrong because using a different origin protocol does not change caching behavior. Option D is wrong because adding more origins does not improve cache hit ratio; it can actually complicate caching.

Answer

Change the origin protocol policy to HTTPS only.

Answer

Add an additional origin server.

Answer

Enable compression for the content.

Question 9

A SysOps administrator is optimizing costs for an Amazon RDS for PostgreSQL instance. The database is used for a reporting application that runs only during business hours (9 AM to 5 PM). Which TWO actions should the administrator take to reduce costs? (Choose TWO.)

Accepted Answer

Stop the database instance during non-business hours.. Correct options: B and C. Option B is correct because using a Multi-AZ deployment is not necessary if the database can be stopped during off-hours; a Single-AZ deployment reduces costs. Option C is correct because stopping the instance during off-hours saves compute costs. Option A is wrong because Reserved Instances require a 1 or 3-year commitment and may not be cost-effective if the database is only used part-time; also, the question asks for cost reduction, not commitment. Option D is wrong because increasing instance size increases cost. Option E is wrong because increasing provisioned IOPS increases cost.

Answer

Move to a larger instance type to handle peaks.

Answer

Increase the provisioned IOPS for better performance.

Answer

Purchase a Reserved Instance for the database instance.

Question 10

A company runs a web application on Amazon EC2 instances behind an Application Load Balancer (ALB). The ALB is configured with a target group that has health checks. The application is experiencing intermittent 503 errors. The SysOps administrator needs to identify the root cause. Which THREE steps should the administrator take to diagnose the issue? (Choose THREE.)

Accepted Answer

Monitor the CloudWatch metric 'HealthyHostCount' for the target group.. Correct options: A, B, D. Option A is correct because checking the ALB access logs can reveal the HTTP response codes and client IPs. Option B is correct because reviewing the target group health check logs helps determine if instances are failing health checks. Option D is correct because CloudWatch metrics like HealthyHostCount can show if the number of healthy instances drops. Option C is wrong because increasing the unhealthy threshold will make the ALB less sensitive, but it does not diagnose the cause. Option E is wrong because checking the security group of the ALB is not directly related to 503 errors; 503 errors typically indicate that there are no healthy targets.

Answer

Check the security group rules for the ALB.

Answer

Increase the unhealthy threshold count in the health check settings.

Question 11

A SysOps administrator wants to reduce the cost of storing infrequently accessed data in Amazon S3. The data must be retained for 10 years for compliance. Which TWO actions should the administrator take? (Choose TWO.)

Accepted Answer

Use S3 Glacier Deep Archive as the destination for the lifecycle transition.. Correct options: B and D. Option B is correct because S3 Lifecycle policies can transition objects to lower-cost storage classes over time. Option D is correct because S3 Glacier Deep Archive is the cheapest storage class for long-term archival. Option A is wrong because S3 Standard is expensive for infrequent access. Option C is wrong because S3 Intelligent-Tiering has a monitoring cost and is best for unknown access patterns. Option E is wrong because enabling versioning increases storage costs due to multiple versions.

Answer

Store the data in S3 Standard.

Answer

Enable S3 Versioning to protect against accidental deletion.

Answer

Use S3 Intelligent-Tiering to automatically optimize costs.

Question 12

A company runs a critical application on a fleet of EC2 instances that process real-time financial transactions. The application requires consistent low latency. The SysOps administrator notices that the application's latency increases periodically due to noisy neighbors. The administrator wants to optimize performance predictability. Which instance type should the administrator choose?

Accepted Answer

Dedicated Instances. Option D is correct because Dedicated Instances run on hardware dedicated to a single customer, eliminating the noisy neighbor effect. Option A is wrong because Burstable Performance Instances (T2, T3) are designed for variable workloads and do not provide consistent performance. Option B is wrong because Spot Instances are not suitable for critical, latency-sensitive applications. Option C is wrong because Reserved Instances provide a discount but do not isolate the instance from other customers.

Answer

Burstable Performance Instances (T3)

Answer

Spot Instances

Answer

Reserved Instances

Question 13

A company hosts a multi-tier web application on AWS. The application consists of an Application Load Balancer (ALB), a fleet of Amazon EC2 instances running in an Auto Scaling group, and an Amazon RDS for MySQL database. The application is accessed by users worldwide. Recently, the company has expanded to new geographic regions, and users in those regions are experiencing high latency. The SysOps administrator is tasked with optimizing performance for global users while keeping costs low. The administrator has already implemented Amazon CloudFront as a CDN for static content. However, dynamic content that requires database queries is still slow. The application's Auto Scaling group is configured with a dynamic scaling policy based on average CPU utilization, but the scaling is not responsive enough during traffic spikes, causing performance degradation. Additionally, the database is a single db.r5.large instance in the us-east-1 region, and all traffic must hit that database, causing high latency for remote users. The administrator needs to propose a comprehensive solution that addresses both compute and database performance issues globally, while considering cost. Which solution is MOST effective?

Accepted Answer

Use Amazon Aurora Global Database to create read replicas in other regions, and configure the Auto Scaling group with a target tracking scaling policy based on request count per target.. Option D is correct because using Aurora Global Database provides a low-latency global read replica for the dynamic content, and using a target tracking scaling policy based on request count per target is more responsive than CPU-based scaling for web applications. Option A is wrong because increasing instance sizes is expensive and does not address global latency. Option B is wrong because adding more instances in a single region does not reduce latency for global users. Option C is wrong because using ElastiCache can reduce database load but does not reduce latency for write operations, and predictive scaling may not be accurate for unpredictable spikes.

Answer

Increase the minimum and maximum size of the Auto Scaling group and use a step scaling policy based on memory utilization.

Answer

Implement Amazon ElastiCache for Redis to cache database queries, and use predictive scaling for the Auto Scaling group.

Answer

Use larger EC2 instances (e.g., c5.2xlarge) for the application tier and provision a Multi-AZ RDS instance for better performance.

Question 14

A company runs a data analytics platform on AWS that processes large datasets using Amazon EMR. The EMR cluster consists of 1 master node (m5.xlarge) and 10 core nodes (r5.2xlarge). The cluster runs batch jobs that last 4 hours each, and the cluster is terminated after each job. The SysOps administrator notices that the cluster takes 15 minutes to provision and bootstrap before each job, which increases the total runtime and cost. The administrator wants to reduce the time and cost associated with cluster provisioning. The data is stored in Amazon S3, and the job scripts are stored in an S3 bucket. The administrator has considered using long-running clusters but is concerned about cost. Which approach should the administrator take to optimize both time and cost?

Accepted Answer

Use EMR Managed Scaling and keep the cluster running between jobs.. Option C is correct because using EMR Managed Scaling with a long-running cluster can reduce provisioning time because the cluster stays up, and managed scaling adjusts the number of core nodes based on the workload, minimizing idle costs. Option A is wrong because using Spot Instances reduces cost but does not reduce provisioning time. Option B is wrong because using a larger master node does not reduce provisioning time. Option D is wrong because using a different file system (Lustre) adds complexity and cost without addressing provisioning time.

Answer

Switch the core nodes to Spot Instances to reduce costs.

Answer

Use a larger master node instance type to speed up bootstrapping.

Answer

Use Amazon FSx for Lustre as the data store to reduce I/O time.

Question 15

A company runs a critical web application on Amazon EC2 instances behind an Application Load Balancer (ALB). The application is deployed in a single AWS region. The company wants to improve performance for users in other regions and reduce the load on the origin servers. The SysOps administrator decides to use Amazon CloudFront. After implementing CloudFront, the administrator notices that the cache hit ratio is very low (around 10%) for the dynamic API responses that are served by the application. These API responses are personalized per user and change frequently. The administrator wants to improve performance for these dynamic requests while still using CloudFront. The application uses cookies for session management and the API responses depend on the user's session. The administrator must ensure that users receive the correct personalized content. Which configuration should the administrator use?

Accepted Answer

Configure CloudFront to use origin shield and enable keep-alive connections to the origin.. Option B is correct because CloudFront can forward cookies to the origin, which allows the origin to generate personalized responses based on the user's session. However, this reduces caching because each user's response is unique. But the question asks to improve performance for dynamic requests. Actually, for dynamic content that is personalized, caching is not effective. Instead, CloudFront can use origin shield or keep-alive connections to improve performance. But the best option among those given is to forward cookies and enable caching based on cookies? Option B says 'Configure CloudFront to forward cookies to the origin and enable caching based on cookies.' That would allow CloudFront to cache responses for different cookie values, but since each user has a unique session, caching would still be inefficient. Option C is better: 'Configure CloudFront to use origin shield and enable keep-alive connections to the origin.' That reduces origin load and improves connection reuse. Option D is about Lambda@Edge for real-time transformation, which could help if the API responses can be assembled at edge. However, the most effective way to improve performance for dynamic requests is to reduce the number of requests hitting the origin by using techniques like origin shield, keep-alive, and possibly Lambda@Edge for caching fragments. But among the options, Option C is the most straightforward. Option A is wrong because disabling caching defeats the purpose. Option B is wrong because caching personalized content based on cookies leads to many cache misses. Option D is possible but complex. The correct answer is C.

Answer

Use Lambda@Edge to generate personalized responses at the edge without forwarding all requests to the origin.

Answer

Configure CloudFront to forward cookies to the origin and enable caching based on cookies.

Answer

Disable caching for the API behavior in CloudFront and enable real-time logs.