Help a team choose a concurrency limit for its own model endpoint. Replay an authored arrival schedule, record queue wait separately from service time, and compare observed percentiles and timeouts across bounded runs. Export a capacity curve and editable operating limit based on measured requests, not model-name lookup tables.
Proposed milestones
- Define an arrival schedule and resource limits
- Measure queue and service time independently
- Compare concurrency settings on the same workload
- Export a reproducible capacity decision
This is an upstream project brief; an implementation and completion tests are not supplied.