Architecture one-pager — [system or capability]
ID: ARCH-[000]
Status: Draft / Reviewed / Accepted
Owner: [Accountable role]
Last reviewed: [YYYY-MM-DD]
Decision horizon: [Date or review trigger]
The decision in one sentence
What must this architecture make possible, and under which constraints?
Business outcomes and users
- Desired outcome and measurable signal:
- Primary human and system consumers:
- Critical business capability / owner:
Scope and boundaries
- In scope:
- Explicitly out of scope:
- Systems of record and ownership:
- Trust / tenant / regulatory boundaries:
Quality attributes (prioritized, measurable)
| Attribute | Target / budget | How validated |
|---|---|---|
| Availability | [target] | [evidence] |
| Latency / freshness | [SLO] | [measurement] |
| Security / privacy | [requirement] | [test / review] |
| Recoverability | [RTO/RPO or policy] | [exercise] |
| Cost | [operating budget] | [metric] |
Design and rationale
- One context diagram, with actors and systems (not vendors as architecture).
- One execution / dataflow diagram, marking boundaries and state.
- Main architectural choices and alternatives; link ADRs.
- Failure modes: timeout, partial commit, duplicate, stale data, unauthorized action.
Contracts and state
- Public APIs / events / schemas:
- Ownership and transaction boundaries:
- Identity and authorization model:
- Idempotency, retries, ordering and compensations:
- Observability, audit and evaluation:
What can go wrong?
| Failure | Detection | Safe behavior | Recovery owner |
|---|---|---|---|
| [mode] | [signal] | [fallback] | [role] |
Decisions still open
- [ ] [Decision and owner]
Review / acceptance
- Business owner:
- Architecture reviewer:
- Security and privacy review:
- Operator / run-team sign-off:
- Evidence and next review trigger: