Examkin

Question

Difficulty: HardPlanning Google Kubernetes Engine (GKE) Cluster Architectures

A healthcare analytics provider is planning a Google Kubernetes Engine (GKE) cluster architecture for a new platform. The environment must support two distinct workloads with specific operational and financial requirements:

1. A high-throughput API gateway service that requires custom Linux kernel parameter modifications (`sysctl` tuning) on the underlying host nodes to optimize socket buffer sizes.
2. An asynchronous, batch processing pipeline that analyzes genomic sequences; this workload is stateless, fault-tolerant, can be interrupted at any time, and must minimize compute expenses.

Which TWO architectural decisions should the cloud engineer incorporate into the GKE design to meet these requirements? (Select TWO.)

  1. Provision a GKE Standard cluster for the API gateway service node pool to support custom OS node configuration and kernel parameter adjustments.Answer
  2. B
    Deploy the high-throughput API gateway service on a GKE Autopilot cluster to eliminate node management while applying custom node OS kernel parameters.
  3. Create a dedicated GKE Standard node pool utilizing Spot Virtual Machines for the genomic batch processing workload to optimize compute costs.Answer
  4. D
    Provision a dedicated Spot Virtual Machine node pool to host the primary stateful relational database of the analytics platform to minimize operational expenses.
  5. E
    Configure the Horizontal Pod Autoscaler (HPA) to add underlying Compute Engine instances to the cluster node pool when genomic batch processing jobs queue up.

Answer

The cloud engineer should select GKE Standard node pools for workloads requiring custom OS sysctl kernel parameter adjustments, and deploy stateless, fault-tolerant genomic batch processing workloads onto dedicated Spot VM node pools for maximum cost savings.
GKE Standard mode provides administrative access to cluster nodes, enabling custom daemonsets or node configuration files to alter kernel sysctl parameters for socket buffer optimization. Additionally, configuring Spot VMs for stateless batch processing delivers substantial cost savings while maintaining architectural resilience, because batch jobs can easily resume if nodes are preempted.

Step-by-Step Solution

1
Evaluate operational requirements for custom host kernel modifications
Identify that GKE Autopilot restricts node-level access and kernel configurations, making GKE Standard necessary for custom sysctl parameters.
GKE Standard retains node pool access and OS configuration capabilities required for low-level network and socket buffer tuning.
2
Evaluate workload fault tolerance for cost optimization options
Identify that stateless, interruptible batch workloads are ideal candidates for Spot VMs.
Spot VMs offer steep discounts for compute capacity in exchange for potential preemption, which aligns perfectly with fault-tolerant batch jobs.
3
Validate distractor architectural flaws
Reject GKE Autopilot for custom kernel configurations, reject Spot VMs for stateful databases, and reject HPA for scaling infrastructure nodes.
Ensures design choices adhere to GCP best practices regarding operational boundaries and autoscaling mechanisms.

Key Concept

GKE Cluster Mode Operational Boundaries and Node Pool Lifecycle Planning
Rate this question

Topics

Planning Google Kubernetes Engine (GKE) Cluster Architectures

More questions from this topic