Examkin

Question

Difficulty: HardPlanning Google Kubernetes Engine (GKE) Cluster Architectures

A financial technology company is planning its Google Cloud infrastructure for two distinct containerized workloads:

1. A stateless transaction routing microservice that requires zero node-management overhead and must automatically scale rapidly during unexpected traffic spikes.
2. A legacy batch calculation engine that requires custom node-level kernel tuning (`sysctl` parameters) and runs non-critical, fault-tolerant data processing tasks.

The platform engineering team needs to optimize operational efficiency and minimize compute costs while adhering to all system constraints. Which TWO architectural decisions should the team implement?

  1. Provision a dedicated GKE Autopilot cluster for the stateless transaction routing microservice to eliminate node provisioning and infrastructure management overhead.Answer
  2. Provision a dedicated GKE Standard cluster with a Spot VM node pool for the legacy batch calculation engine to allow custom sysctl kernel configurations and minimize compute costs.Answer
  3. C
    Deploy the legacy batch calculation engine on a GKE Autopilot cluster to allow Google to automatically manage and tune node kernel sysctl parameters.
  4. D
    Configure the stateless transaction routing microservice to run on Spot VMs in a GKE Standard cluster to ensure high availability at minimum baseline cost.
  5. E
    Use the Horizontal Pod Autoscaler (HPA) within a single GKE Autopilot cluster to dynamically apply node-level kernel sysctl parameters when the batch workload scales up.

Answer

The correct architectural decisions are: 1) Deploying the stateless transaction routing microservice on a GKE Autopilot cluster to eliminate node management, and 2) Deploying the legacy batch calculation engine on a GKE Standard cluster with Spot VMs to allow custom sysctl kernel settings while minimizing compute costs.
GKE Autopilot provides a fully managed environment where Google handles node provisioning and management, making it optimal for stateless microservices that require low operational overhead. GKE Standard is necessary for workloads that require node-level customizations such as custom sysctl kernel settings. Provisioning a GKE Standard node pool with Spot VMs for fault-tolerant batch workloads optimizes compute costs significantly.

Step-by-Step Solution

1
Analyze the requirements for the stateless transaction routing microservice.
The microservice needs zero node-management overhead and fast autoscaling.
GKE Autopilot fully manages node provisioning, scaling, and maintenance, making it the ideal operational choice for stateless workloads requiring low overhead.
2
Analyze the requirements for the legacy batch calculation engine.
The workload requires custom node-level sysctl kernel tuning and is fault-tolerant.
Custom node kernel parameters require GKE Standard because Autopilot restricts system-level OS modifications. Because the workload is fault-tolerant, utilizing Spot VMs minimizes compute costs.
3
Evaluate why combining these workloads in a single GKE Autopilot or Standard configuration fails constraints.
Autopilot blocks custom node sysctl settings, and Spot VMs are unsafe for critical transaction routing services.
A multi-cluster approach isolates node-level OS requirements while matching cost and availability profiles to each workload.

Key Concept

Selecting between GKE Autopilot and GKE Standard cluster architectures based on operational boundaries (custom node OS/kernel configurations) and workload tolerance (Spot VMs vs. high availability).
Estimated Time:2m 0s
Rate this question

Topics

Planning Google Kubernetes Engine (GKE) Cluster Architectures

More questions from this topic