Non-functional Requirements Examples
Make NFRs explicit and testable — performance, availability, security, recoverability, cost — attached to the architecture they constrain.
What belongs in a non-functional requirements
Good NFRs are numbers, not adjectives: p95 latency < 300ms, 99.9% monthly availability, RPO 15 minutes, data classified and encrypted at rest, a monthly cost ceiling. Every one should be testable, owned, and tied to the architecture decision that satisfies it — vague NFRs are where projects quietly fail.
How Architext builds it
Architext keeps NFRs in the same model as the diagram and decisions, and compiles them into the solution-architecture document — so the targets, the design that meets them, and the doc reviewers sign off on stay consistent.
Document, export, and share
Once the document is right, compile the full solution-architecture document, generate an audience-tailored slide deck, export to PNG, SVG, Mermaid, or PowerPoint, or share a read-only link — all from the same model.
FAQ
- What are common NFR categories?
- Performance, scalability, availability, security, recoverability (RPO/RTO), observability, compliance, and cost.
- What makes an NFR good?
- It's measurable and testable — "p95 under 300ms at 1k RPS", not "fast". If you can't test it, it's a wish.
- Where should NFRs live?
- With the architecture, not in a separate wiki — in Architext they're part of the model and the generated SAD.
Non-functional Requirements Examples
Free to start — paste your code or describe your system and watch the diagram build itself.
Open Architext