Breaking Changes: 1.13 → 2.0
OpenMetadata 2.0 is a major release. It retires the thread-backed collaboration model, replaces the Explore experience, moves LLM and embedding configuration out of the search block, adds a database-backed session store, and reshapes several API contracts. This section documents every change that can break an existing client, script, deployment or user workflow — broken out by component so you only have to read the parts that apply to you.How to read this section
Every entry is classified so you can triage quickly.
Each entry states what changed, who is affected and what to do.
The changes most likely to break you
Deprecation notices
These are not 2.0 breaking changes, but they change what you should build on now.Browse by component
API & Schema Contracts
Removed endpoints, changed request and response shapes, tightened validation, enum and default changes.
Discovery & Search (Explore)
The redesigned Explore page, the browse-and-filter query bar, changed URL parameters, facet scoping and staged ranking.
Collaboration
Task redesign, removal of the Suggestions API, standalone Announcements, the Activity Stream and alert filters.
Data Governance
Policies and new task operations, glossary and ontology, classification, governance workflows and the Policy Agent.
Data Quality & Observability
Incidents as Tasks, inlined incident status, auto-close, multi-status filters and the removed Data Insights DQ module.
Lineage
FQN-based edge APIs, batch hydration, time-window traversal semantics and new lineage change events.
Ingestion & Connectors
pipelineStatuses, Databricks Pipeline auth, Great Expectations 1.x, log-stream SSE and new connectors.Applications & Automation
Search Indexing and RDF app config, Data Insights modules, MCP tool contracts and CSV background jobs.
Platform, Config & Security
llmConfiguration, session management, openmetadata.yaml changes and the 2.0.0 database migration.UI, Personas & Customization
Design-system consolidation, App Mode, persona preferences and removed UI components.
Upgrade checklist
1
Inventory your API clients
Search scripts, dashboards, Terraform and CI jobs for
/v1/suggestions, /v1/feed,
search_after=, pipelineStatuses and recreateIndex.2
Diff your openmetadata.yaml against the 2.0 template
The
elasticsearch.naturalLanguageSearch.* provider blocks no longer exist. Port them to
llmConfiguration before you upgrade.3
Record your application configurations
SearchIndexingApplication, DataInsightsApplication, RdfIndexApp and McpApplication
configurations are all rewritten by the migration.4
Check entity names for newly-invalid characters
>, " and ASCII control characters are rejected by entityName validation in 2.0.5
Back up the database and run the upgrade
Expect these data migrations: suggestions →
task_entity, thread tasks → task_entity,
system feed threads → activity_stream, announcements → announcement_entity, and
thread_entity renamed to thread_entity_legacy.6
Allow a full search reindex
Reindexing in 2.0 always recreates the index. There is no incremental mode, and the ranking
configuration is backfilled into
searchSettings during the migration.7
Verify after upgrade
Check Explore result ordering, alert subscriptions that relied on partial FQN matching,
Snowflake/Databricks/Unity Catalog
policyAgentConfig defaults, and whether any user routinely
exceeds five concurrent sessions.Version scope
A few entries reference capabilities that ship in Collate rather than OpenMetadata OSS — AI
Governance Studio, the Policy Agent, Context Center and AI Mode. They are documented here because
their schemas and REST namespaces are part of the 2.0 specification and appear in the OpenAPI
surface either way.