Examkin

Soru

Zorluk: OrtaPlanning and Assessing Cloud Compute Solutions

A media publishing platform is migrating two backend workloads to Google Cloud with distinct operational and cost constraints:

1. Workload 1: A stateless HTTP microservice packaged as a container image that experiences unpredictable web traffic throughout the day. It requires automatic scaling down to zero instances when idle and zero operational maintenance for virtual machine instances or Kubernetes node pools.
2. Workload 2: A nightly batch processing job that runs for several hours. The pipeline is fully fault-tolerant and saves checkpoint progress to Cloud Storage every 5 minutes. The team's primary priority is to minimize compute costs for this batch execution.

Which TWO compute platform strategies should the Cloud Engineer recommend to satisfy these requirements? (Select TWO choices.)

  1. Deploy Workload 1 on Cloud Run to leverage serverless container execution with automatic scaling to zero and no infrastructure management.Cevap
  2. Deploy Workload 2 on Compute Engine using Spot Virtual Machines (VMs) to significantly reduce compute costs for the fault-tolerant batch job.Cevap
  3. C
    Deploy Workload 1 on Cloud Functions by rewriting the microservice into single-purpose event handlers to support multi-container orchestration.
  4. D
    Deploy Workload 2 on GKE Autopilot using Spot nodes to host a non-fault-tolerant stateful database requiring persistent in-memory session cache.
  5. E
    Deploy Workload 1 on Compute Engine Managed Instance Groups (MIGs) with standard VMs to achieve zero infrastructure overhead.

Cevap

The optimal solution is to deploy Workload 1 on Cloud Run and deploy Workload 2 on Compute Engine using Spot VMs.
Deploying Workload 1 on Cloud Run satisfies all requirements for containerized microservices needing scale-to-zero capability and zero infrastructure maintenance. Deploying Workload 2 on Compute Engine using Spot VMs minimizes compute costs significantly while taking advantage of the job's checkpointing resilience against VM preemption.

Adım Adım Çözüm

1
Evaluate Workload 1 requirements (stateless container, unpredictable traffic, scale-to-zero, zero infrastructure management).
Identify Cloud Run as the best fit because it runs arbitrary container images, scales automatically to zero when no HTTP requests arrive, and requires no node or OS management.
Matching workload architecture to the minimal operational overhead serverless compute service.
2
Evaluate Workload 2 requirements (batch processing, nightly run, checkpointing to Cloud Storage, cost minimization as top priority).
Identify Compute Engine Spot VMs as the best fit because Spot VMs provide heavy cost discounts (60-90%) suitable for fault-tolerant batch workloads.
Choosing compute purchasing models based on workload fault tolerance and cost targets.

Anahtar Kavram

Selecting GCP compute options based on workload statefulness, operational overhead, and purchasing models.
Tahmini Süre:1m 30s
Bu soruyu puanla

Konular

Planning and Assessing Cloud Compute Solutions

Bu konudan daha fazla soru