A Kubernetes cluster runs multiple microservices. One image-rendering service is CPU-heavy; others are sized for n1-standard types. How do you optimize resource efficiency?

The correct option optimizes resource efficiency by matching workloads to appropriate machine types. CPU-heavy services, like image rendering, benefit from compute-optimized nodes (e.g., C2 machine types in GCP) which provide higher CPU-to-memory ratios and better performance for CPU-bound tasks. General-purpose nodes (e.g., n1-standard) are suitable for most other microservices. This approach prevents the CPU-heavy service from monopolizing resources on general-purpose nodes, which could starve other services, and avoids over-provisioning expensive compute-optimized resources for less demanding workloads. Assigning higher pod priority (first incorrect option) only affects scheduling order, not the underlying node resources, so it won't optimize resource efficiency for a CPU-heavy workload. The third option incorrectly assigns the CPU-heavy service to general-purpose nodes and the others to compute-optimized, which is inefficient. Configuring resource requests (fourth option) is a good practice for all pods but doesn't address the fundamental need for specialized hardware for a CPU-intensive service to achieve optimal efficiency.