Examkin

Question

Difficulty: MediumPlanning Google Kubernetes Engine (GKE) Cluster Architectures

An online video processing platform is designing a Google Kubernetes Engine (GKE) cluster architecture to support two distinct workloads:

1. A real-time web API serving client requests where the operations team requires fully managed infrastructure with zero node provisioning or cluster node maintenance overhead.
2. A stateless frame-rendering batch processing pipeline that is fault-tolerant and requires significant cost reduction on compute resources.

Which TWO architectural choices should you combine to meet these workload requirements while minimizing operational complexity and compute expense?

  1. Deploy a GKE Autopilot cluster for the real-time web API workload.Answer
  2. Provision a node pool composed of Spot VMs in GKE Standard for the stateless batch rendering workload.Answer
  3. C
    Deploy a GKE Standard cluster with fixed custom node pools to eliminate node management overhead for the web API.
  4. D
    Provision a dedicated Spot VM node pool to host the platform's persistent stateful database workloads.
  5. E
    Configure the Horizontal Pod Autoscaler (HPA) to add physical Compute Engine instances when batch rendering demand spikes.

Answer

Deploy a GKE Autopilot cluster for the real-time web API workload, and provision a Spot VM node pool for the stateless batch rendering workload.
Combining GKE Autopilot for the web API workload and a Spot VM node pool for the batch workload satisfies all requirements. Autopilot offloads all node management and infrastructure maintenance to Google Cloud, matching the zero-node-management constraint for the web API. Spot VMs offer up to 60-90% savings for stateless, fault-tolerant batch workloads where preemption does not cause data loss.

Step-by-Step Solution

1
Analyze the web API requirements
GKE Autopilot is selected because Google fully manages the underlying nodes, scaling, security hardening, and OS maintenance.
The requirement explicitly asks for zero node management and maintenance overhead.
2
Analyze the stateless batch rendering workload requirements
Spot VMs are selected to drastically reduce compute costs.
Spot VMs are heavily discounted compute instances ideal for stateless, fault-tolerant batch processing that can sustain preemptions.

Key Concept

GKE Cluster Architectural Modes and Node Pool Selection Strategy
Rate this question

Topics

Planning Google Kubernetes Engine (GKE) Cluster Architectures

More questions from this topic