CPU metrics lag demand by minutes because traffic arrives as queue depth. How should autoscaling react to events?

Utilization-based scaling is the wrong sensor when work arrives as backlog. The design questions are threshold math against drain rate, reaction-time bounds from polling, and why lag-driven fleets oscillate without generous targets and idempotent consumers.

Grounded in researched DevOps, SRE and platform engineering interview loops, written to a senior-engineer editorial bar, and never padded to hit a word count.