Examkin

Soru

Zorluk: Çok zorPlanning Google Kubernetes Engine (GKE) Cluster Architectures

An enterprise logistics company is redesigning its Google Cloud container infrastructure to support two distinct workload profiles. The first profile consists of stateless web APIs that experience highly variable traffic, where the operations team requires zero node management responsibility and pod-level SLA guarantees. The second profile consists of a batch analytics engine executing fault-tolerant, short-lived jobs that require deep customization of worker node kernel parameters (sysctl) and maximum cost reduction. Which TWO GKE cluster architecture decisions should the cloud architect implement to meet these requirements? (Select TWO.)

  1. Provision a GKE Autopilot cluster for the stateless web APIs to offload node configuration, maintenance, and node autoscaling to Google.Cevap
  2. Provision a GKE Standard cluster with a Spot VM node pool configured with custom sysctl settings for the batch analytics engine.Cevap
  3. C
    Provision a single GKE Autopilot cluster and configure custom node daemonsets with elevated host permissions to apply sysctl settings for the batch analytics engine.
  4. D
    Provision a GKE Standard cluster with Spot VM node pools for the stateless web APIs to eliminate node management overhead.
  5. E
    Provision a GKE Autopilot cluster using Spot pods with custom sysctl flags for both workloads to minimize operational costs.

Cevap

The correct decisions are to deploy a GKE Autopilot cluster for the stateless web APIs and a GKE Standard cluster with a Spot VM node pool featuring custom sysctl configurations for the batch analytics engine.
Deploying GKE Autopilot for the stateless web APIs fulfills the zero node management overhead requirement, as Google manages the underlying nodes. Deploying GKE Standard with a Spot VM node pool for the batch analytics engine permits custom sysctl kernel configurations while minimizing compute costs for fault-tolerant workloads.

Adım Adım Çözüm

1
Analyze the operational requirements for the stateless web APIs.
The requirements demand zero node management overhead and automated node maintenance. GKE Autopilot completely manages underlying infrastructure, making it the ideal fit.
GKE Autopilot abstracts node management away from the operator while guaranteeing pod-level SLAs.
2
Analyze the technical requirements for the batch analytics engine.
The batch analytics engine requires custom sysctl kernel configurations and maximum cost optimization for fault-tolerant workloads.
Custom node-level sysctl configurations require GKE Standard, while Spot VMs supply deep discounts for fault-tolerant compute.
3
Evaluate architectural separation.
Combining both requirements into Autopilot fails because Autopilot blocks custom node kernel modifications. Standard is required for node OS level customization.
Operational boundaries in Autopilot enforce strict security constraints preventing node-level privileged modifications.

Anahtar Kavram

Selecting between GKE Autopilot and GKE Standard based on node OS customization boundaries and choosing Spot VMs for cost-sensitive, fault-tolerant workloads.
Tahmini Süre:3m 0s
Bu soruyu puanla

Konular

Planning Google Kubernetes Engine (GKE) Cluster Architectures

Bu konudan daha fazla soru