NEWS, 24 AUG 2026, read 13 Nov 2026
OpenMetadata 2.0.0 revamps data quality defaults and adds certification intake forms
OpenMetadata published 2.0.0 on GitHub on 24 Aug 2026. The notes group the data quality changes under a revamp and add governance intake forms, a Context Center that replaces Knowledge Center, and MCP changes.
Per the notes, the Profiler now defaults to Dynamic Sampling, which the project says reduces query cost and run time on large tables. Cardinality distribution is no longer collected on every run, and existing profiler configs are preserved. Custom Intake Forms can be defined per entity type for workflows such as asset certification and ownership assignment. The MCP server is enabled by default after upgrade, and read tools return a nextCursor value for stable paging.
Two consequences for a readiness program follow. Any rule that depends on distinct-value counts must now add cardinality back to the profiler configuration, or the check silently loses its input. And sampled profiles are estimates, which matters for a completeness gate on small tables. The certification form gives a place to record who approved a table, but the notes do not describe an expiry for that approval.
More readiness news
- Monte Carlo adds an Analytics API that reports data health through its MCP server4 Nov 2026
- Databricks makes data quality monitoring anomaly detection generally available29 Oct 2026
- Qlik introduces Semantic Models for data products and plans MCP tool support28 Oct 2026
- DataHub 1.8.0 nests data products and adds a Monte Carlo ingestion source22 Oct 2026
Sources
- github.com/open-metadata/OpenMetadata/releases/tag/2.0.0-release, read 13 Nov 2026
Last reviewed 13 Nov 2026.