Scale and performance

Large estates can contain tens of thousands of objects and generate high event volume during assessment import, readiness evaluation, and wave execution. Size the deployment with representative source inventory, event, and artifact workloads rather than object counts alone.

Required benchmark shapes

• 10k and 50k objects across multiple estates and waves;
• assessment import and dependency projection time;
• steady-state incremental hydration latency;
• concurrent writer plus projection restart/recovery;
• wave approval/cutover fan-out event volume;
• evidence export size and artifact-store latency;
• readiness evaluation and stale-certificate scans.

Record the application build, store configuration, dataset generator or source snapshot, concurrency, run duration, and latency percentiles with each result. Set an accepted operating envelope for the project, alert before it is exceeded, and repeat the benchmark after storage, projection, policy, or evidence-provider changes.