Capacity Forecasting Laboratory
Project workload from measured CPU and I/O buckets, simulate FIFO queues and compare fixed capacity with delayed replica scaling.
Application guideVerifying your session…
Using Capacity Forecasting Laboratory
Compare service capacity and queue growth using supplied workload observations.
Seasonal demand profiles · Multi-resource saturation · Replica scaling and queues
- Inputs and working data
- Uniform UTC buckets with arrivals, CPU seconds and I/O operations, seasonal demand assumptions and bounded replica policies.
- Working in the editor
- Measurement import, selectable forecast window, policy editor, comparative queue chart and interval table.
- What can I produce?
- Policy comparison, backlog and resource saturation curves, individual projected intervals and portable forecast ZIP.
- What are the limits?
- 5,000 historical buckets, 336 seasonal phases, 672 forecast buckets, 16 policies. Continuous aggregate service, linear per-season demand growth and initially empty queue.
- Example
- Ingest CPU-limited workload, compare one replica with lagged scale-out and reconcile final backlog and served requests.