Developer workflow
Data contracts in CI
Stop learning about breaking changes from a broken dashboard. Validate contracts and preview blast radius on every pull request — before merge.
What CI should gate
Spec validation
YAML pipeline contracts validate against schema before they can register.
Impact / blast radius
Diff the change against lineage — who breaks if this mart shifts?
Consumer compatibility
Consumers declare freshness and row expectations; producers must stay compatible.
Passport awareness
See which shareable trust warrants would revoke if the change ships broken.
PR flow
- 1
Open a PR
Engineer changes a model, DAG, or pipeline spec in Git.
- 2
CI runs impact
GitHub Action validates the contract and calls blast-radius analysis.
- 3
Review with context
PR comment lists downstream consumers and Passport risk — merge only when green.
Runtime + CI, same contract
CI gates prevent bad merges. Freshness and KPI checks catch silent failures after deploy. Pipeline Passports only stay issued when both sides agree the data is trustworthy.
Data contracts CI FAQ
Gates that engineers actually run.
What is a data contract in CI?+
A data contract is an enforceable promise about schema, freshness, volume, or semantics. In CI, that means every PR that changes a pipeline runs validation and impact analysis so breaking changes fail the check before production.
How is this different from dbt model contracts?+
dbt model contracts enforce column shape at build time — excellent for schema. DataXPipe adds consumer-declared staleness/row expectations, Passport status, and PR blast-radius comments across producers and downstream consumers.
What does DataXPipe post on a pull request?+
The GitHub Action calls impact analysis (POST /specs/impact) and comments with downstream datasets, affected consumer contracts, and Passport risk — so reviewers see blast radius without opening a separate tool.
Put blast radius on the next PR
Free tier includes 2 pipelines. Wire the GitHub Action after guided setup.