Examkin

Soru

Zorluk: Çok zorPlanning and Assessing Cloud Compute Solutions

An organization is evaluating Google Cloud compute solutions for two newly developed services with distinct operational profiles:

1. Service Alpha: A stateless HTTP webhook receiver that experiences unpredictable request bursts. Single HTTP request execution times range from a few seconds up to 45 minutes. The team requires zero infrastructure/node management and automatic scaling down to zero when idle.
2. Service Beta: A batch analytics engine that performs large-scale computations overnight. The application is stateless, includes checkpointing logic to resume execution if interrupted, and must be deployed with maximum cost efficiency.

Which TWO compute architectural configurations should the cloud engineer select to meet the requirements for Service Alpha and Service Beta? (Select TWO.)

  1. Deploy Service Alpha on Cloud Run with a request timeout configured to 45 minutes.Cevap
  2. Deploy Service Beta on Compute Engine instances using Spot VMs within a Managed Instance Group.Cevap
  3. C
    Deploy Service Alpha on Cloud Functions (2nd gen) utilizing default function execution timeouts.
  4. D
    Deploy Service Beta on Compute Engine using standard On-Demand VMs to prevent node preemption during nightly batch runs.
  5. E
    Deploy Service Alpha on GKE Standard node pools to achieve zero operational and node management overhead.

Cevap

The optimal architecture combines Cloud Run with an extended request timeout for Service Alpha, and Compute Engine Spot VMs in a Managed Instance Group for Service Beta.
The combination of deploying Service Alpha on Cloud Run (configured with a 45-minute request timeout) and Service Beta on Compute Engine Spot VMs provides the exact alignment with technical and financial requirements. Cloud Run manages infrastructure entirely, scales to zero, and allows HTTP request timeouts up to 60 minutes. Spot VMs provide maximum cost savings for batch jobs that are fault-tolerant and use checkpointing.

Adım Adım Çözüm

1
Assess the requirements for Service Alpha (HTTP webhook, stateless, 45-minute execution, scale to zero, zero node management).
Identify that Cloud Run supports stateless HTTP container workloads, scales to zero, eliminates node management, and allows configuring request timeouts up to 60 minutes.
Serverless containers fit the operational requirements without requiring VM or cluster lifecycle management.
2
Assess the requirements for Service Beta (stateless batch processing, checkpointing capability, extreme cost sensitivity).
Identify that Compute Engine Spot VMs offer massive cost reductions (60-91% off list price) and are fully suitable for workloads designed to handle node preemptions via checkpointing.
Fault-tolerant batch jobs should leverage excess GCP compute capacity via Spot VMs rather than paying full price for standard compute instances.
3
Evaluate and eliminate distractors based on operational constraints and service boundaries.
Eliminate Cloud Functions (due to default execution limits and framework constraints), standard On-Demand VMs (due to unnecessary cost for fault-tolerant workloads), and GKE Standard (due to manual node management overhead).
Selecting appropriate GCP compute resources requires balancing operational overhead, execution limits, and pricing models.

Anahtar Kavram

Selecting GCP Compute options based on workload statefulness, operational overhead, execution timeouts, and cost optimization trade-offs.
Bu soruyu puanla

Konular

Planning and Assessing Cloud Compute Solutions

Bu konudan daha fazla soru