14 — Performance Standards

Status: Ratified Version: 1.0.0 Applies to: All components, tests/performance/, capacity planning, scaling


1. Performance Goals

Goal Target
CLI cold start < 1.5s to first command
API p95 latency (read) < 150ms
API p95 latency (write/ingest) < 300ms
Tool step dispatch overhead < 5ms
Canonical event ingest ≥ 10,000 events/s (single node, batch)
Finding triage query (1M findings) < 100ms p95
Mission with 1,000 targets completes within profile constraints
Concurrent missions ≥ 200 active (single node)
Horizontal scale-out ≥ 1,000 concurrent targets across workers

2. Memory


3. CPU


4. Concurrency & Parallelism


5. Caching & Hot-Path Optimization


6. Large-Target Support

Target sizing model (approximate per engagement):

Class Targets Notes
small < 100 single node default
medium 100–10k worker fleet; batch tool runs
large 10k–1M partition by scope; chunked tool invocation; streaming ingest
extreme > 1M distributed workers; sharded DB; aggregated reporting

Mechanisms:


7. Database Optimization


8. Reporting Performance


9. Performance Testing


10. Operational SLAs


11. References