| name | data-contract-design |
| description | Defines producer-consumer contracts for schema, semantics, quality, compatibility and operations. Use when formalising API, event, or dataset agreements between teams. |
Data Contract Design
When to use
Use where a stable producer-consumer boundary exists for tables, files, APIs, events or data products.
Objective
Produce a practical, concise, traceable architecture artefact that a coding agent can use to guide implementation or review.
Procedure
- Identify producer and consumers.
- Define interface and scope.
- Define schema, semantics and constraints.
- Define quality, freshness and completeness obligations.
- Define compatibility/versioning.
- Define classification and allowed use.
- Define contract tests.
- Define incident/deprecation process.
Required outputs
- Producer/consumer list
- Schema and semantic definitions
- Quality/freshness obligations
- Compatibility policy
- Contract tests
- Deprecation process
Best-practice alignment
Apply DAMA-DMBOK2-style separation of data governance, architecture, modelling, security, integration/interoperability, master/reference data, metadata and quality. For cloud/shared data, apply CDMC-style expectations: ownership, classification, entitlement/access evidence, lineage/provenance, lifecycle/retention, quality controls and auditable evidence.
Align producer-consumer contracts with the Open Data Contract Standard (ODCS) where a portable YAML/JSON contract is appropriate: fundamentals, schema, quality rules, team/ownership, roles, SLA, support channels and infrastructure. Prefer contract-first design, semantic versioning for breaking changes, contracts co-located with the data product, and executable validation in CI (for example Data Contract CLI).
Quality checks
- Contract has syntax and semantics.
- Breaking changes are defined.
- Tests are specified.
- Ownership and support are explicit.
Avoid
Do not define schemas without semantics or change management.
Do not treat contracts as documentation-only; enforce them with tests and pipeline checks.
References
Verification
