Amazon AWS Certified Solutions Architect - Associate SAA-C03 Bundle

Use Lambda behind API Gateway for bounded synchronous request handling would help if a synchronous web request handler completes in seconds and should not require a running web fleet. Lambda plus API Gateway offers request-driven compute and HTTPS routing. That is not the symptom described here: thousands of independent already-containerized research tasks must run in parallel with isolated retries. The matching response is Submit AWS Batch array jobs using job-level retry settings and appropriate compute, because array jobs partition a large batch into separately scheduled units.

For the alternative failure a portable container app needs task-level scaling without cluster host operations, the control Use ECS services on Fargate is relevant: fargate supplies managed container compute capacity. Here the evidence instead is thousands of independent already-containerized research tasks must run in parallel with isolated retries. The selected technical mechanism is Submit AWS Batch array jobs using job-level retry settings and appropriate compute: array jobs partition a large batch into separately scheduled units.

Use EC2 instances or an appropriate container service rather than Lambda solves a different operational problem where a service runs a continuous daemon for days with custom OS dependencies. Lambda execution duration is bounded. It does not directly address thousands of independent already-containerized research tasks must run in parallel with isolated retries. The case calls for Submit AWS Batch array jobs using job-level retry settings and appropriate compute, since array jobs partition a large batch into separately scheduled units.

Choose memory-optimized instances sized for measured memory footprint would help if an in-memory analytics job needs very large RAM relative to vCPU. Memory-optimized families offer high memory-to-CPU ratios. That is not the symptom described here: a numeric simulation saturates CPU but uses little memory. The matching response is Choose compute-optimized EC2 instance families based on profiling, because compute-optimized types emphasize CPU resources.

For the alternative failure machine-learning inference requires GPUs not available on existing general instances, the control Choose appropriate accelerated EC2 instances for the inference framework is relevant: gPU acceleration supports supported parallel model operations. Here the evidence instead is a numeric simulation saturates CPU but uses little memory. The selected technical mechanism is Choose compute-optimized EC2 instance families based on profiling: compute-optimized types emphasize CPU resources.

The competing option Choose EC2 types with sufficient enhanced networking and bandwidth would be useful when a service needs a high network packet rate but instance bandwidth caps throughput. Instance networking specifications constrain maximum throughput. The present issue is a numeric simulation saturates CPU but uses little memory; it is more directly addressed by Choose compute-optimized EC2 instance families based on profiling. Compute-optimized types emphasize CPU resources.

For the alternative failure a service needs a high network packet rate but instance bandwidth caps throughput, the control Choose EC2 types with sufficient enhanced networking and bandwidth is relevant: instance networking specifications constrain maximum throughput. Here the evidence instead is an in-memory analytics job needs very large RAM relative to vCPU. The selected technical mechanism is Choose memory-optimized instances sized for measured memory footprint: memory-optimized families offer high memory-to-CPU ratios.

Choose appropriate accelerated EC2 instances for the inference framework solves a different operational problem where machine-learning inference requires GPUs not available on existing general instances. GPU acceleration supports supported parallel model operations. It does not directly address an in-memory analytics job needs very large RAM relative to vCPU. The case calls for Choose memory-optimized instances sized for measured memory footprint, since memory-optimized families offer high memory-to-CPU ratios.

The option Choose compute-optimized EC2 instance families based on profiling addresses a separate condition: a numeric simulation saturates CPU but uses little memory. Compute-optimized types emphasize CPU resources. The actual requirement is an in-memory analytics job needs very large RAM relative to vCPU. Choose memory-optimized instances sized for measured memory footprint instead: Memory-optimized families offer high memory-to-CPU ratios.

Choose EC2 types with sufficient enhanced networking and bandwidth solves a different operational problem where a service needs a high network packet rate but instance bandwidth caps throughput. Instance networking specifications constrain maximum throughput. It does not directly address machine-learning inference requires GPUs not available on existing general instances. The case calls for Choose appropriate accelerated EC2 instances for the inference framework, since gPU acceleration supports supported parallel model operations.

The competing option Evaluate general-purpose EC2 families with benchmark data would be useful when a balanced web server needs predictable CPU, RAM and networking. General-purpose instances balance common resource dimensions. The present issue is machine-learning inference requires GPUs not available on existing general instances; it is more directly addressed by Choose appropriate accelerated EC2 instances for the inference framework. GPU acceleration supports supported parallel model operations.

The option Choose compute-optimized EC2 instance families based on profiling addresses a separate condition: a numeric simulation saturates CPU but uses little memory. Compute-optimized types emphasize CPU resources. The actual requirement is machine-learning inference requires GPUs not available on existing general instances. Choose appropriate accelerated EC2 instances for the inference framework instead: GPU acceleration supports supported parallel model operations.

The competing option Choose compute-optimized EC2 instance families based on profiling would be useful when a numeric simulation saturates CPU but uses little memory. Compute-optimized types emphasize CPU resources. The present issue is a balanced web server needs predictable CPU, RAM and networking; it is more directly addressed by Evaluate general-purpose EC2 families with benchmark data. General-purpose instances balance common resource dimensions.

The option Choose appropriate accelerated EC2 instances for the inference framework addresses a separate condition: machine-learning inference requires GPUs not available on existing general instances. GPU acceleration supports supported parallel model operations. The actual requirement is a balanced web server needs predictable CPU, RAM and networking. Recommended corrective action: Evaluate general-purpose EC2 families with benchmark data. General-purpose instances balance common resource dimensions.

For the alternative failure an in-memory analytics job needs very large RAM relative to vCPU, the control Choose memory-optimized instances sized for measured memory footprint is relevant: memory-optimized families offer high memory-to-CPU ratios. Here the evidence instead is a balanced web server needs predictable CPU, RAM and networking. The selected technical mechanism is Evaluate general-purpose EC2 families with benchmark data: general-purpose instances balance common resource dimensions.

The option Choose compute-optimized EC2 instance families based on profiling addresses a separate condition: a numeric simulation saturates CPU but uses little memory. Compute-optimized types emphasize CPU resources. The actual requirement is a service needs a high network packet rate but instance bandwidth caps throughput. Choose EC2 types with sufficient enhanced networking and bandwidth instead: Instance networking specifications constrain maximum throughput.

For the alternative failure a balanced web server needs predictable CPU, RAM and networking, the control Evaluate general-purpose EC2 families with benchmark data is relevant: general-purpose instances balance common resource dimensions. Here the evidence instead is a service needs a high network packet rate but instance bandwidth caps throughput. The selected technical mechanism is Choose EC2 types with sufficient enhanced networking and bandwidth: instance networking specifications constrain maximum throughput.

Choose appropriate accelerated EC2 instances for the inference framework solves a different operational problem where machine-learning inference requires GPUs not available on existing general instances. GPU acceleration supports supported parallel model operations. It does not directly address a service needs a high network packet rate but instance bandwidth caps throughput. The case calls for Choose EC2 types with sufficient enhanced networking and bandwidth, since instance networking specifications constrain maximum throughput.

Use Gateway Load Balancer and compatible appliance routing would help if third-party stateful inspection devices must see transparent bidirectional flows. GWLB provides transparent appliance traffic distribution. That is not the symptom described here: requests need routing by URL path to different container target groups. The matching response is Use an ALB with path-based listener rules, because aLB routes at HTTP Layer 7.

For the alternative failure a target group continuously marks healthy instances unhealthy due to the wrong port, the control Correct target-group health-check port/protocol and threshold settings is relevant: health checks control available targets. Here the evidence instead is requests need routing by URL path to different container target groups. The selected technical mechanism is Use an ALB with path-based listener rules: aLB routes at HTTP Layer 7.

The competing option Choose NLB with the matching transport protocol listener would be useful when TLS/TCP clients require fast stable IP-based ingress. NLB supports Layer-4 connection routing. The present issue is requests need routing by URL path to different container target groups; it is more directly addressed by Use an ALB with path-based listener rules. ALB routes at HTTP Layer 7.

For the alternative failure a target group continuously marks healthy instances unhealthy due to the wrong port, the control Correct target-group health-check port/protocol and threshold settings is relevant: health checks control available targets. Here the evidence instead is TLS/TCP clients require fast stable IP-based ingress. The selected technical mechanism is Choose NLB with the matching transport protocol listener: nLB supports Layer-4 connection routing.

Use Gateway Load Balancer and compatible appliance routing solves a different operational problem where third-party stateful inspection devices must see transparent bidirectional flows. GWLB provides transparent appliance traffic distribution. It does not directly address TLS/TCP clients require fast stable IP-based ingress. The case calls for Choose NLB with the matching transport protocol listener, since nLB supports Layer-4 connection routing.

The option Use ALB capabilities that support WebSocket HTTP connections addresses a separate condition: a custom HTTP service needs WebSocket support and cookie-related routing. ALB handles HTTP upgrade flows. The actual requirement is TLS/TCP clients require fast stable IP-based ingress. Choose NLB with the matching transport protocol listener instead: NLB supports Layer-4 connection routing.

Correct target-group health-check port/protocol and threshold settings solves a different operational problem where a target group continuously marks healthy instances unhealthy due to the wrong port. Health checks control available targets. It does not directly address third-party stateful inspection devices must see transparent bidirectional flows. The case calls for Use Gateway Load Balancer and compatible appliance routing, since gWLB provides transparent appliance traffic distribution.

The competing option Choose NLB with the matching transport protocol listener would be useful when TLS/TCP clients require fast stable IP-based ingress. NLB supports Layer-4 connection routing. The present issue is third-party stateful inspection devices must see transparent bidirectional flows; it is more directly addressed by Use Gateway Load Balancer and compatible appliance routing. GWLB provides transparent appliance traffic distribution.

The option Use an ALB with path-based listener rules addresses a separate condition: requests need routing by URL path to different container target groups. ALB routes at HTTP Layer 7. The actual requirement is third-party stateful inspection devices must see transparent bidirectional flows. Recommended corrective action: Use Gateway Load Balancer and compatible appliance routing. GWLB provides transparent appliance traffic distribution.

The competing option Use Gateway Load Balancer and compatible appliance routing would be useful when third-party stateful inspection devices must see transparent bidirectional flows. GWLB provides transparent appliance traffic distribution. The present issue is a custom HTTP service needs WebSocket support and cookie-related routing; it is more directly addressed by Use ALB capabilities that support WebSocket HTTP connections. ALB handles HTTP upgrade flows.

Correct target-group health-check port/protocol and threshold settings would help if a target group continuously marks healthy instances unhealthy due to the wrong port. Health checks control available targets. That is not the symptom described here: a custom HTTP service needs WebSocket support and cookie-related routing. The matching response is Use ALB capabilities that support WebSocket HTTP connections, because aLB handles HTTP upgrade flows.

For the alternative failure requests need routing by URL path to different container target groups, the control Use an ALB with path-based listener rules is relevant: aLB routes at HTTP Layer 7. Here the evidence instead is a custom HTTP service needs WebSocket support and cookie-related routing. The selected technical mechanism is Use ALB capabilities that support WebSocket HTTP connections: aLB handles HTTP upgrade flows.

Use an ALB with path-based listener rules would help if requests need routing by URL path to different container target groups. ALB routes at HTTP Layer 7. That is not the symptom described here: a target group continuously marks healthy instances unhealthy due to the wrong port. The matching response is Correct target-group health-check port/protocol and threshold settings, because health checks control available targets.

For the alternative failure TLS/TCP clients require fast stable IP-based ingress, the control Choose NLB with the matching transport protocol listener is relevant: nLB supports Layer-4 connection routing. Here the evidence instead is a target group continuously marks healthy instances unhealthy due to the wrong port. The selected technical mechanism is Correct target-group health-check port/protocol and threshold settings: health checks control available targets.

Use ALB capabilities that support WebSocket HTTP connections solves a different operational problem where a custom HTTP service needs WebSocket support and cookie-related routing. ALB handles HTTP upgrade flows. It does not directly address a target group continuously marks healthy instances unhealthy due to the wrong port. The case calls for Correct target-group health-check port/protocol and threshold settings, since health checks control available targets.