Examkin

Soru

Zorluk: OrtaOptimizing Business Processes through FinOps and Cloud Cost Management

A digital agriculture enterprise operates a Google Cloud environment to ingest daily IoT sensor data and run seasonal machine learning (ML) crop yield models. The environment features a predictable, 24/7 baseline web and API service alongside bursty, highly variable ML batch processing workloads. Currently, all resources are billed at standard on-demand rates, resulting in high monthly spend. The executive team has mandated a FinOps governance strategy to optimize costs without risking budget lock-in for unpredictable workloads. Which architecture and purchasing strategy should the Cloud Architect recommend?

  1. Acquire spend-based Flexible Committed Use Discounts (CUDs) to cover the steady-state baseline compute usage, leverage Spot VMs for the transient ML batch workloads, and implement mandatory resource labeling tied to BigQuery Cloud Billing exports for granular cost allocation.Cevap
  2. B
    Purchase 3-year resource-based standard Committed Use Discounts covering the maximum peak capacity required during seasonal ML batch model training across all regional projects.
  3. C
    Migrate all transient ML batch workloads and baseline API services into a dedicated regional Google Kubernetes Engine (GKE) cluster with static node pools to eliminate on-demand overhead.
  4. D
    Assign the finance team primitive IAM Owner roles across all GCP project hierarchies so they can directly terminate unutilized Compute Engine instances and reconfigure project billing accounts manually.

Cevap

Acquire spend-based Flexible Committed Use Discounts (CUDs) to cover the steady-state baseline compute usage, leverage Spot VMs for the transient ML batch workloads, and implement mandatory resource labeling tied to BigQuery Cloud Billing exports for granular cost allocation.
Combining Flexible Committed Use Discounts for predictable baseline services with Spot VMs for stateless batch ML jobs provides maximum financial savings while retaining architectural agility. Linking mandatory labels to BigQuery billing exports forms the foundation of cloud financial accountability (FinOps).

Adım Adım Çözüm

1
Analyze workload profile and separate baseline capacity from dynamic peak usage.
Identified steady-state 24/7 web/API services and fault-tolerant, highly variable seasonal ML batch processing.
Different workload patterns require distinct cost optimization mechanisms to maximize savings without over-committing.
2
Select appropriate commitment models and compute purchasing tiers.
Applied Flexible CUDs to baseline services for predictable discount coverage across regions/families, and assigned Spot VMs to ML batch processing.
Flexible CUDs secure baseline savings while Spot VMs dramatically lower cost for stateless, fault-tolerant batch workloads.
3
Establish FinOps accountability and governance mechanisms.
Enforced resource labels and enabled detailed billing exports to BigQuery.
Labeling and automated export analytics enable unit-economics tracking and showback/chargeback governance across teams.

Anahtar Kavram

FinOps Cloud Cost Optimization and Governance
Bu soruyu puanla

Konular

Optimizing Business Processes through FinOps and Cloud Cost Management