[WIP] v1.0.0 Roadmap

[jjk-g]

P0

• [Loadgen] Multi-model or LoRA support and traffic splitting
• [Datagen] Shared prefix datagen incorrect prompt length #230

P1

• [Datagen] Support different input / output distribution for different stages (helpful for autoscaling)
• [Metrics] GPU utilization / other hardware metrics from Prometheus
• [Metrics] SLO support and conformance of specific latency SLOs

P2

[jjk-g]

Notes from 1/22/26 meeting

• Added functionality to analysis module (multiple runs on same graph, benchmark results summary/analysis)
• Review test plan with @huaxig and code coverage. + code coverage presubmit.
• Review crucial bug fixes and add (concurrency issue)
• Image + Helm chart publishing