Which approach best supports autoscaling decisions?

Which approach best supports autoscaling decisions?

• A. Relying on a single static threshold
• B. Exposing comprehensive metrics and monitoring
• C. Disabling telemetry
• D. Rerouting all traffic manually

Answer: (B)

Understanding autoscaling decisions hinges on having visibility into how the system is performing. When you expose comprehensive metrics and monitoring, the autoscaler can compare real workload signals—such as CPU and memory usage, request rate, latency, error rate, and custom application metrics—against desired targets. This data-driven approach lets the system decide when to add or remove replicas or adjust resources in near real time, matching capacity to demand and reducing the risk of over- or under-provisioning. Relying on a single static threshold is inflexible and often misreads dynamic workloads. It can cause the autoscaler to react too slowly or too aggressively when traffic patterns shift, leading to thrashing or resource waste. Disabling telemetry eliminates the visibility needed to make informed scaling choices, leaving you blind to actual demand. Manually rerouting traffic bypasses automation entirely, preventing timely scaling in response to changing load and undermining the purpose of autoscaling.

Understanding autoscaling decisions hinges on having visibility into how the system is performing. When you expose comprehensive metrics and monitoring, the autoscaler can compare real workload signals—such as CPU and memory usage, request rate, latency, error rate, and custom application metrics—against desired targets. This data-driven approach lets the system decide when to add or remove replicas or adjust resources in near real time, matching capacity to demand and reducing the risk of over- or under-provisioning.

Relying on a single static threshold is inflexible and often misreads dynamic workloads. It can cause the autoscaler to react too slowly or too aggressively when traffic patterns shift, leading to thrashing or resource waste. Disabling telemetry eliminates the visibility needed to make informed scaling choices, leaving you blind to actual demand. Manually rerouting traffic bypasses automation entirely, preventing timely scaling in response to changing load and undermining the purpose of autoscaling.