A media streaming platform is designing a Google Kubernetes Engine (GKE) cluster deployment to support two distinct workloads: an internal management portal that requires custom Linux kernel sysctl modifications on cluster worker nodes, and a batch telemetry processing job that runs nightly and can tolerate unexpected pod terminations. Which TWO cluster configuration choices should the cloud engineer implement to meet these requirements?
- Provision a GKE Standard cluster to support custom kernel parameter configuration on node instances.Answer
- Configure a dedicated node pool utilizing Spot VMs for the nightly batch processing pipeline.Answer
- CDeploy the management portal to a GKE Autopilot cluster to reduce node configuration overhead.
- DHost the internal management portal on a Spot VM node pool to lower baseline compute expenses.
- EConfigure Horizontal Pod Autoscaler (HPA) to scale the number of Compute Engine worker nodes during peak batch jobs.
Answer
The correct architectural decisions are to provision a GKE Standard cluster (to allow custom node kernel sysctl parameters) and to configure a dedicated Spot VM node pool for the fault-tolerant nightly batch workload.
GKE Standard mode allows administrators to customize node configurations, such as sysctl settings, which are restricted in GKE Autopilot. Additionally, using Spot VMs in a dedicated node pool for batch workloads maximizes cost reduction while honoring the fault-tolerant nature of the nightly processing jobs.
Step-by-Step Solution
Key Concept
Selecting between GKE Standard vs Autopilot based on node OS customization constraints, and choosing appropriate node pool lifecycle types (Spot vs On-Demand) based on workload fault tolerance.
Estimated Time:1m 30s