Examkin

Soru

Zorluk: OrtaDesigning Infrastructure for Business Requirements and Cost Optimization

A financial analytics firm processes daily transaction reconciliation jobs that run for only 3 hours every midnight, requiring high CPU for processing but zero compute capacity for the remaining 21 hours of the day. Additionally, their operational relational dataset is approximately 400 GB400\text{ GB} in size with standard regional availability requirements. The company wants to minimize baseline compute and database costs while ensuring automated scaling and minimal management overhead. Which TWO architectural decisions should you recommend to meet these business requirements? (Select TWO.)

  1. Containerize the daily reconciliation workload and execute it using Cloud Run jobs triggered by Cloud Scheduler.Cevap
  2. Provision a Cloud SQL for PostgreSQL instance with automatic storage increase enabled to store the relational dataset.Cevap
  3. C
    Deploy a Google Kubernetes Engine (GKE) Autopilot cluster with a fixed minimum pool of 10 nodes to handle batch execution readiness.
  4. D
    Migrate the relational database to a multi-region Cloud Spanner instance to ensure maximum availability and seamless scale.
  5. E
    Purchase a 3-year Committed Use Discount (CUD) for Compute Engine instances dedicated to running the midnight batch script.

Cevap

The organization should containerize the batch workload using Cloud Run jobs triggered by Cloud Scheduler, and host the regional 400 GB400\text{ GB} relational database on Cloud SQL.
The solution minimizes operational costs by aligning compute usage strictly with execution time using serverless Cloud Run jobs (which incur zero charges when idle) and right-sizing the relational database using Cloud SQL, which meets regional 400 GB400\text{ GB} storage needs efficiently without the node overhead of Cloud Spanner.

Adım Adım Çözüm

1
Analyze compute cost requirements for short-duration batch workloads
Identified that compute runs only 3 hours per day (12.5%12.5\% duty cycle), making serverless scale-to-zero compute (Cloud Run jobs) optimal over continuous VMs or baseline Kubernetes clusters.
Running dedicated VMs or GKE nodes continuously results in paying for 21 hours of unused compute time daily.
2
Evaluate database requirements against data scale and SLA
The dataset is 400 GB400\text{ GB} within a single region, which falls within standard regional relational database boundaries.
Cloud SQL delivers required performance and high availability at significantly lower cost than multi-region Cloud Spanner.

Anahtar Kavram

Selecting serverless compute for bursty/scheduled batch jobs and cost-effective regional relational storage for moderate data volumes.
Tahmini Süre:1m 30s
Bu soruyu puanla

Konular

Designing Infrastructure for Business Requirements and Cost Optimization