Examkin

Soru

Zorluk: ZorPlanning Google Kubernetes Engine (GKE) Cluster Architectures

A digital publishing platform is planning its Google Kubernetes Engine (GKE) cluster architecture to support two core backend workloads in a single environment:

1. A real-time content rendering engine that requires custom sysctl kernel parameter modifications on the node host operating system to optimize socket buffers.
2. A background image thumbnail generation pipeline that processes asynchronous worker queues, where individual jobs are stateless, idempotent, and can be safely re-queued if interrupted.

The engineering team wants to satisfy both workload requirements while maximizing cost efficiency and avoiding unsupportable configurations. Which GKE cluster architecture and node pool strategy should the team implement?

  1. Provision a GKE Standard cluster. Configure a dedicated node pool with custom Linux sysctl configuration parameters applied to the underlying node configuration for the content rendering workload, and configure a secondary node pool leveraging Spot VMs for the thumbnail generation pipeline.Cevap
  2. B
    Provision a GKE Autopilot cluster. Configure custom sysctl parameters directly within the pod deployment specification for the content rendering workload, and deploy the thumbnail generation workload to default Autopilot compute classes.
  3. C
    Provision a GKE Standard cluster. Configure a dedicated node pool using Spot VMs for the real-time content rendering workload to reduce core infrastructure costs, and configure a standard persistent node pool for the background thumbnail generation pipeline.
  4. D
    Provision a GKE Standard cluster. Enable the Horizontal Pod Autoscaler (HPA) to automatically provision additional Compute Engine node instances whenever cluster CPU utilization exceeds 80%.

Cevap

Provision a GKE Standard cluster with a custom sysctl-configured node pool for rendering and a Spot VM node pool for background processing.
The option specifying GKE Standard with two targeted node pools is correct because GKE Standard is mandatory when custom Linux sysctl kernel settings must be configured on cluster nodes. Furthermore, leveraging Spot VMs for the stateless, idempotent image processing workload delivers maximum cost savings without compromising the availability of the real-time rendering service.

Adım Adım Çözüm

1
Analyze host OS kernel custom parameter requirements
GKE Autopilot locks down node OS settings and disallows custom kernel sysctl parameters; GKE Standard is required.
When workloads demand low-level Linux kernel modifications on the node, GKE Standard provides direct control over node pool configuration.
2
Analyze fault tolerance and cost requirements for background processing
Spot VMs are ideal for stateless, idempotent worker queues.
Spot VMs offer significant compute cost savings (up to 80%) for workloads that can tolerate node preemptions without causing data loss.
3
Synthesize the optimal cluster architecture
A GKE Standard cluster with two specialized node pools fulfills all constraints.
Combining a custom-configured Standard node pool for sensitive workloads and a Spot VM node pool for batch processing balances operational capability with cost optimization.

Anahtar Kavram

GKE Autopilot vs. Standard Operational Boundaries & Node Pool Selection
Tahmini Süre:2m 0s
Bu soruyu puanla

Konular

Planning Google Kubernetes Engine (GKE) Cluster Architectures

Bu konudan daha fazla soru