DataEngBytes 2023 - PER-T1-01 - Chad Sanderson
summary
Claims that YAML contracts in Git plus CI/CD GitHub Actions enforcing tiers turn data quality from reactive toil into preventative control by validating schema, semantics and SLAs before deploy. Implementation at Convoy compares developer migration output to contract, tags downstream consumers in PRs, and integrates with Amundsen catalog and monitoring; tier 0 breaks the build, tier 1 alerts producer and consumer, tier 2 alerts consumer. Problem framing cites monthly outages costing millions and that >75% of teams rank data quality top investment.
AI-written summary · about
chapters (12)
- 0:00AI data quality crisis
- 2:07Poor data breaks AI
- 8:11Low data quality causes
- 11:21Why conventional fixes fail
- 13:48Preventative solution requirements
- 14:29What is a data contract
- 19:43Implementing contracts with YAML
- 21:30Enforcing contracts in CI/CD
- 24:42Where to enforce contracts
- 26:26Contracts with catalogs and monitoring
- 27:09Changing culture with contracts
- 29:17Key takeaways and Q&A
related talks
41mDataEngBytes 2023 - BRI-T1-01 - Chad Sanderson
28mDataEngBytes 2023 - SYD-T1-03 - Chad Sanderson
21mDataEngBytes 2023 - SYD-T3-04 - Abhinav Goyal
27mLive in the Data Wild West: The Data Contracts Sheriff
31mData Reliability Engineering: Secret Of Our Success
24mDataEngBytes 2023 - MEL-T2-02 - Vikas Rajput
32mDataEngBytes 2023 - BRI-T1-10 Rishu Saxena & John Cosgrove
24mDataEngBytes 2023 - MEL-T1-05 - Izzudin Hafiz
