Amazon AWS Certified Solutions Architect - Associate SAA-C03 Bundle

Configure target tracking and safe scale-in policies would help if containers are held at maximum task count after a weekend burst. Target tracking removes capacity as demand falls. That is not the symptom described here: tests create large EC2 instances and never terminate them. The matching response is Automate resource expiration with tags and lifecycle rules, because ephemeral environment cleanup reduces orphaned compute cost.

For the alternative failure the workload spikes at 9 a.m. daily and scaling arrives late, the control Configure scheduled scale-out ahead of the spike is relevant: predictable capacity can be provisioned before demand. Here the evidence instead is tests create large EC2 instances and never terminate them. The selected technical mechanism is Automate resource expiration with tags and lifecycle rules: ephemeral environment cleanup reduces orphaned compute cost.

Scope provisioned concurrency to critical aliases and active hours solves a different operational problem where a Lambda API has strict low latency but provisioned concurrency is enabled for every function 24/7. Prepared concurrency has ongoing costs even without invocations. It does not directly address tests create large EC2 instances and never terminate them. The case calls for Automate resource expiration with tags and lifecycle rules, since ephemeral environment cleanup reduces orphaned compute cost.

Evaluate committed accelerated EC2 capacity after benchmarking would help if a custom native workload requires fixed GPUs continuously. Persistent GPU usage can benefit from targeted reservations. That is not the symptom described here: ten thousand independent simulations need compute for only two hours per day. The matching response is Use AWS Batch with elastic compute environments and appropriate Spot support, because batch can scale tasks down when the queue drains.

For the alternative failure video processing needs purpose-built managed encoding rather than custom GPU fleet maintenance, the control Use AWS Elemental MediaConvert for supported transcoding jobs is relevant: managed conversion avoids maintaining dedicated workers. Here the evidence instead is ten thousand independent simulations need compute for only two hours per day. The selected technical mechanism is Use AWS Batch with elastic compute environments and appropriate Spot support: batch can scale tasks down when the queue drains.

The competing option Use Lambda with appropriate memory and concurrency would be useful when a workload executes 30-second image transformations upon upload. Short event processing fits serverless billing. The present issue is ten thousand independent simulations need compute for only two hours per day; it is more directly addressed by Use AWS Batch with elastic compute environments and appropriate Spot support. Batch can scale tasks down when the queue drains.

Use ECS on Fargate with task autoscaling solves a different operational problem where a containerized application must run a long-lived API without Kubernetes management. Fargate removes host administration and scales tasks. It does not directly address video processing needs purpose-built managed encoding rather than custom GPU fleet maintenance. The case calls for Use AWS Elemental MediaConvert for supported transcoding jobs, since managed conversion avoids maintaining dedicated workers.

The competing option Evaluate committed accelerated EC2 capacity after benchmarking would be useful when a custom native workload requires fixed GPUs continuously. Persistent GPU usage can benefit from targeted reservations. The present issue is video processing needs purpose-built managed encoding rather than custom GPU fleet maintenance; it is more directly addressed by Use AWS Elemental MediaConvert for supported transcoding jobs. Managed conversion avoids maintaining dedicated workers.

The option Use AWS Batch with elastic compute environments and appropriate Spot support addresses a separate condition: ten thousand independent simulations need compute for only two hours per day. Batch can scale tasks down when the queue drains. The actual requirement is video processing needs purpose-built managed encoding rather than custom GPU fleet maintenance. Recommended corrective action: Use AWS Elemental MediaConvert for supported transcoding jobs. Managed conversion avoids maintaining dedicated workers.

The competing option Evaluate committed accelerated EC2 capacity after benchmarking would be useful when a custom native workload requires fixed GPUs continuously. Persistent GPU usage can benefit from targeted reservations. The present issue is a workload executes 30-second image transformations upon upload; it is more directly addressed by Use Lambda with appropriate memory and concurrency. Short event processing fits serverless billing.

The option Use ECS on Fargate with task autoscaling addresses a separate condition: a containerized application must run a long-lived API without Kubernetes management. Fargate removes host administration and scales tasks. The actual requirement is a workload executes 30-second image transformations upon upload. Recommended corrective action: Use Lambda with appropriate memory and concurrency. Short event processing fits serverless billing.

Use AWS Elemental MediaConvert for supported transcoding jobs would help if video processing needs purpose-built managed encoding rather than custom GPU fleet maintenance. Managed conversion avoids maintaining dedicated workers. That is not the symptom described here: a workload executes 30-second image transformations upon upload. The matching response is Use Lambda with appropriate memory and concurrency, because short event processing fits serverless billing.

The competing option Use Lambda with appropriate memory and concurrency would be useful when a workload executes 30-second image transformations upon upload. Short event processing fits serverless billing. The present issue is a containerized application must run a long-lived API without Kubernetes management; it is more directly addressed by Use ECS on Fargate with task autoscaling. Fargate removes host administration and scales tasks.

Evaluate committed accelerated EC2 capacity after benchmarking would help if a custom native workload requires fixed GPUs continuously. Persistent GPU usage can benefit from targeted reservations. That is not the symptom described here: a containerized application must run a long-lived API without Kubernetes management. The matching response is Use ECS on Fargate with task autoscaling, because fargate removes host administration and scales tasks.

For the alternative failure video processing needs purpose-built managed encoding rather than custom GPU fleet maintenance, the control Use AWS Elemental MediaConvert for supported transcoding jobs is relevant: managed conversion avoids maintaining dedicated workers. Here the evidence instead is a containerized application must run a long-lived API without Kubernetes management. The selected technical mechanism is Use ECS on Fargate with task autoscaling: fargate removes host administration and scales tasks.

The option Use AWS Batch with elastic compute environments and appropriate Spot support addresses a separate condition: ten thousand independent simulations need compute for only two hours per day. Batch can scale tasks down when the queue drains. The actual requirement is a custom native workload requires fixed GPUs continuously. Recommended corrective action: Evaluate committed accelerated EC2 capacity after benchmarking. Persistent GPU usage can benefit from targeted reservations.

For the alternative failure a workload executes 30-second image transformations upon upload, the control Use Lambda with appropriate memory and concurrency is relevant: short event processing fits serverless billing. Here the evidence instead is a custom native workload requires fixed GPUs continuously. The selected technical mechanism is Evaluate committed accelerated EC2 capacity after benchmarking: persistent GPU usage can benefit from targeted reservations.

Use AWS Elemental MediaConvert for supported transcoding jobs solves a different operational problem where video processing needs purpose-built managed encoding rather than custom GPU fleet maintenance. Managed conversion avoids maintaining dedicated workers. It does not directly address a custom native workload requires fixed GPUs continuously. The case calls for Evaluate committed accelerated EC2 capacity after benchmarking, since persistent GPU usage can benefit from targeted reservations.

Right-size to memory-optimized EC2 if memory pressure dominates would help if a large in-memory analytical cache spills to disk on undersized general-purpose hosts. Memory-optimized instances reduce spill overhead. That is not the symptom described here: a CPU-bound data transform runs on an expensive memory-optimized instance. The matching response is Move to compute-optimized instances after validating memory needs, because cPU-centric families avoid paying for unused RAM.

For the alternative failure oversized EBS bandwidth is purchased though storage throughput is far lower, the control Select an instance size matching measured network and EBS bandwidth is relevant: instance type choice should reflect actual I/O requirements. Here the evidence instead is a CPU-bound data transform runs on an expensive memory-optimized instance. The selected technical mechanism is Move to compute-optimized instances after validating memory needs: cPU-centric families avoid paying for unused RAM.

Use general-purpose EC2 instances sized to observed demand solves a different operational problem where a general web tier has balanced modest CPU and memory requirements. Balanced families suit common mixed workloads. It does not directly address a CPU-bound data transform runs on an expensive memory-optimized instance. The case calls for Move to compute-optimized instances after validating memory needs, since cPU-centric families avoid paying for unused RAM.

Move to compute-optimized instances after validating memory needs solves a different operational problem where a CPU-bound data transform runs on an expensive memory-optimized instance. CPU-centric families avoid paying for unused RAM. It does not directly address a large in-memory analytical cache spills to disk on undersized general-purpose hosts. The case calls for Right-size to memory-optimized EC2 if memory pressure dominates, since memory-optimized instances reduce spill overhead.

The competing option Replace it with suitable CPU instances or Lambda would be useful when a GPU instance runs a simple CPU-only scheduled script. GPU capacity is unnecessary for CPU-only tasks. The present issue is a large in-memory analytical cache spills to disk on undersized general-purpose hosts; it is more directly addressed by Right-size to memory-optimized EC2 if memory pressure dominates. Memory-optimized instances reduce spill overhead.

The option Use general-purpose EC2 instances sized to observed demand addresses a separate condition: a general web tier has balanced modest CPU and memory requirements. Balanced families suit common mixed workloads. The actual requirement is a large in-memory analytical cache spills to disk on undersized general-purpose hosts. Recommended corrective action: Right-size to memory-optimized EC2 if memory pressure dominates. Memory-optimized instances reduce spill overhead.

The competing option Select an instance size matching measured network and EBS bandwidth would be useful when oversized EBS bandwidth is purchased though storage throughput is far lower. Instance type choice should reflect actual I/O requirements. The present issue is a general web tier has balanced modest CPU and memory requirements; it is more directly addressed by Use general-purpose EC2 instances sized to observed demand. Balanced families suit common mixed workloads.

The option Replace it with suitable CPU instances or Lambda addresses a separate condition: a GPU instance runs a simple CPU-only scheduled script. GPU capacity is unnecessary for CPU-only tasks. The actual requirement is a general web tier has balanced modest CPU and memory requirements. Recommended corrective action: Use general-purpose EC2 instances sized to observed demand. Balanced families suit common mixed workloads.

Move to compute-optimized instances after validating memory needs would help if a CPU-bound data transform runs on an expensive memory-optimized instance. CPU-centric families avoid paying for unused RAM. That is not the symptom described here: a general web tier has balanced modest CPU and memory requirements. The matching response is Use general-purpose EC2 instances sized to observed demand, because balanced families suit common mixed workloads.

The competing option Right-size to memory-optimized EC2 if memory pressure dominates would be useful when a large in-memory analytical cache spills to disk on undersized general-purpose hosts. Memory-optimized instances reduce spill overhead. The present issue is a GPU instance runs a simple CPU-only scheduled script; it is more directly addressed by Replace it with suitable CPU instances or Lambda. GPU capacity is unnecessary for CPU-only tasks.

The option Select an instance size matching measured network and EBS bandwidth addresses a separate condition: oversized EBS bandwidth is purchased though storage throughput is far lower. Instance type choice should reflect actual I/O requirements. The actual requirement is a GPU instance runs a simple CPU-only scheduled script. Recommended corrective action: Replace it with suitable CPU instances or Lambda. GPU capacity is unnecessary for CPU-only tasks.

For the alternative failure a CPU-bound data transform runs on an expensive memory-optimized instance, the control Move to compute-optimized instances after validating memory needs is relevant: cPU-centric families avoid paying for unused RAM. Here the evidence instead is a GPU instance runs a simple CPU-only scheduled script. The selected technical mechanism is Replace it with suitable CPU instances or Lambda: gPU capacity is unnecessary for CPU-only tasks.

Right-size to memory-optimized EC2 if memory pressure dominates would help if a large in-memory analytical cache spills to disk on undersized general-purpose hosts. Memory-optimized instances reduce spill overhead. That is not the symptom described here: oversized EBS bandwidth is purchased though storage throughput is far lower. The matching response is Select an instance size matching measured network and EBS bandwidth, because instance type choice should reflect actual I/O requirements.

For the alternative failure a GPU instance runs a simple CPU-only scheduled script, the control Replace it with suitable CPU instances or Lambda is relevant: gPU capacity is unnecessary for CPU-only tasks. Here the evidence instead is oversized EBS bandwidth is purchased though storage throughput is far lower. The selected technical mechanism is Select an instance size matching measured network and EBS bandwidth: instance type choice should reflect actual I/O requirements.

Use general-purpose EC2 instances sized to observed demand solves a different operational problem where a general web tier has balanced modest CPU and memory requirements. Balanced families suit common mixed workloads. It does not directly address oversized EBS bandwidth is purchased though storage throughput is far lower. The case calls for Select an instance size matching measured network and EBS bandwidth, since instance type choice should reflect actual I/O requirements.