Advanced · Capacity

Plan capacity for a service that must survive losing a zone at peak

A service runs 12 instances across three zones at 80% CPU during peak, autoscaling on CPU with a maximum of 16. Leadership wants to cut infrastructure spend and points at the idle capacity outside peak hours. Produce a capacity plan.

The fix that looks right

Trap

Raising the CPU target from 70% to 85% and lowering the minimum instance count, which reduces spend immediately and shows no impact in normal operation for months. It works because the headroom being removed is exactly the headroom that is never used until it is — the zone loss, the traffic spike, the deploy at peak. The cost saving is real and the risk it buys is invisible until a single event converts it into an outage, at which point the connection between the two is six months old and nobody makes it.