Overview
Government cloud demand is spiky around service deadlines and campaigns. This runbook forecasts capacity from historical telemetry so teams scale ahead of demand instead of reacting to it.
Steps
- Pull CPU/memory/storage telemetry per service.
- AI models seasonality and event-driven spikes to forecast the next N weeks.
- Flag services trending toward their limits with a recommended headroom.
- Feed the forecast into the capacity-planning review.
Impact
Fewer incidents from capacity crunches, and less over-provisioning spend.