Release 2026.03.00
Featured
.NET 10, Elastic Search and RabbitMQ Upgrade
Upgraded the CluedIn platform to .NET 10.
Upgraded Elastic Search from 7.17.29 to 9.5.1.
Upgraded RabbitMQ from 3.7 to 3.13.7.
This brings several overall improvements and security fixes along with many new features.
Any custom-built extensions for CluedIn should be upgraded to match CluedIn’s .NET 10 platform reference versions. In most cases, this would involve updating CluedIn NuGet package dependencies to version 5.0.2 and updating the target framework for the extension to .NET 10.
There will be a data migration step required when upgrading to this version. Please refer to our documentation for further details.
Custom Jobs and Job Monitoring
The new Job and Monitoring functionality provides greater visibility and control over background operations across CluedIn. Jobs can now be viewed and managed directly through the UI, allowing users to monitor execution, review job history, and cancel running jobs when necessary.
Users can also create custom jobs, with or without a schedule, to automate a variety of operations. These include reprocessing rules, performing bulk actions such as reprocessing records, providing greater flexibility in how and when these activities are performed.
In addition, system jobs are now managed through the same centralized framework, providing a unified experience for monitoring and managing both system and custom jobs. This brings greater transparency, flexibility, and control to background processing across the platform.
AI Agent Improvement
AI Agents now include a Copilot experience that lets users configure and refine agents using natural conversation. Copilot can guide users through setting up jobs, selecting data, interpreting results, and progressing work from testing into production.
Enricher Multi Match
Enrichers now support multiple candidate matches for a golden record, providing greater flexibility and control over enrichment results.
The new Matches experience allows users to compare potential matches, approve or reject candidates, and configure automatic selection using rules or AI. This streamlines the enrichment process while giving users greater confidence in the accuracy and quality of their data.
Global Data Model Enhancements
The Global Data Model has been enhanced to provide a clearer, more intuitive view of how data is structured and connected across CluedIn.
The improved interface makes it easier to explore business domains and understand relationships, with new capabilities to size domains by record count, filter by data source, and display domains even when no relationships exist.
These improvements provide greater visibility into your data landscape, making it easier to understand how information is organized and connected.
New Notifications
The Notification Center has been redesigned to provide a clearer, more organized experience for tracking and managing notifications across CluedIn.
Notifications now update dynamically rather than creating duplicates, with support for multi-step progress tracking, attachments, and single-use actions. Improved filtering, search, and grouping capabilities make it easier to find relevant information, track ongoing activities, and manage related notifications efficiently.
Release Notes
Cluedin
Features
AI & Copilot
- CluedIn Copilot now uses “business domain” instead of “entity type” when creating or listing business domain information (#51604)
- Azure AI Foundry deployments now support all three base URL formats (
services.ai.azure.com, cognitiveservices.azure.com, and openai.azure.com), with Claude models automatically detected and routed through the Anthropic Messages API (#56823)
- Copilot search skills now include a lightweight Simple Search option and support paginated results for navigating large result sets (#57394)
- Copilot now supports AI Agents and AI Jobs (#57662)
- AI endpoint cache duration can now be configured (#58440)
- A new “AI Agent Chat” feature flag, available in beta and disabled by default, controls the AI agent chat panel, AI agent review suggestions for uploaded datasets, and agent-to-job chat delegation
- Copilot dataset onboarding now supports mapping-first plans and chat visualizations (#58555)
Rules & Survivorship
- Survivorship rules can now select the minimum or maximum value (#56368)
- Survivorship rules can now select a data part based on a condition (#57785)
- Invalid lookup rules now include operators for filtering by the specific reason a lookup failed
- The Copy Value action is now available in golden record rules (#55283)
- Rule audit logs now record who started or cancelled reprocessing (#58409)
PowerFx
- PowerFx can now set
Revision on data parts
- PowerFx now supports
Exists and Not Exists operators across supported data types (#57414)
- PowerFx now supports regular expressions through the
Match and IsMatch functions (#58029)
- PowerFx can now load entity relationships using
LoadOutgoingRelationships and LoadIncomingRelationships (#58044)
Enrichers
- Enricher mapping can now be configured to hide the golden record primary identifier toggle (#56462)
- Enrichments can now be tested to evaluate their impact on golden records before being applied (#55137)
- A new Enricher Matches screen provides a consolidated view for reviewing multiple enrichment results (#57217)
Deduplication
- Deduplication projects can now be automated with selected normalizations and a limited number of entities in scope (#58083)
- Deduplication projects now show when matches were generated, both in the review header and in a sortable “Last generation” column on the projects list (#57628)
- Deduplication project audit logs now record who enabled or stopped automation
- A new API endpoint returns the deduplication groups an entity belongs to across projects (#57627)
Access Control
- Access control policy actions can now exclude specific vocabularies and vocabulary keys from included items (#57540)
Security
- Password requirements have been strengthened, with a new endpoint allowing the UI to display the current requirements (#57464)
- Invalid passwords now return clearer errors when creating users or organizations (#58054)
Jobs & Scheduling
- New APIs support the Jobs screen and retrieving progress for multiple jobs (#58121)
- Custom jobs can now be created and managed through the Custom Jobs system (#58123)
- Job history retention can now cap the number of completed jobs retained per organization, defaulting to 100,000 and configurable through
DistributedJobs.JobRetentionMaxJobs (#58573)
- Archive metrics, high-resolution metric trimming, stream log retention, stream idle checks, and record deduplication by entity code are now managed as Custom Jobs, with their schedules and applicable settings configurable through the UI (#58573)
Data Management
- Configuration can now be exported and imported through new utilities and API controllers
- Organizations can now be reset to their default state or deleted (#52918)
- The global data model has been improved (#58466)
- Several legacy data-processing behaviours now default to disabled, including author population,
/Author edge creation, last-changed metadata population, temporal edge creation, tag edge creation, and ignoring minimum or maximum date values
- The solution has been updated to .NET 10 (#56314)
- SQL database health checks now use more efficient connectivity checks and verify database size usage (#57534)
- Multi-layer relationship queries have been optimized to improve performance (#57767)
Fixes
Search & Lookups
- Lookup “Search All” queries now work correctly when source values are numeric (#57123)
- Updated lookup key display names are now reflected correctly (#56150)
- Value pickers now load existing values more efficiently for fields with high cardinality (#56717)
AI & Copilot
- Local AI connection errors are no longer repeatedly logged when Ollama has not been configured; the self-hosted AI base URL must now be explicitly configured when using a local Ollama instance (#57355)
- Copilot no longer encounters a loop error caused by
tool_calls exceptions (#58449)
- AI Agents now reuse existing deduplication projects when working with matches (#58240)
- AI agent summaries no longer require unnecessary full-detail loads, improving efficiency (#58609)
- Data shard health checks now enforce a retry delay when a health check takes longer than the configured interval (#57534)
- Locking scopes are now disposed of correctly, preventing potential connection pool starvation (#57632)
- Entity serialization no longer performs an Elasticsearch query for vocabulary key metadata on every request, resolving a performance regression that could add 200–350 ms to affected API calls (#57656)
CluedIn.MicroServices
Features
Enrichers
- A new Enricher Matches screen provides a consolidated view for reviewing multiple enrichment results (#57217)
Datasets & Ingestion
- New events are available for created and processed dataset uploads (#58183)
- Data source pages now load more efficiently
- Processing statistics on dataset preview and processing tabs now load more efficiently
- Dataset mappings now load more efficiently
- Dataset content queries have been optimized for improved performance
- Large dataset content views now perform more efficiently through cached sort-key discovery, fewer Elasticsearch requests, lighter permission queries, safer pagination, capped page sizes, and improved slow-query diagnostics
- Dataset edge queries have been optimized for improved performance
- Duplicate detection is now faster
- Mapping now strips unsupported Unicode characters from data already ingested into Elasticsearch (#58156)
Mapping
- Mapping previews are now significantly faster and cross-tenant error handling has also been improved (#58209)
- ID batching now processes a maximum of 1,000 IDs at a time, improving efficiency and reliability
- Indexes have been introduced for IDs awaiting processing to improve query performance
- Ingestion log queries have been optimized with indexes on dataset ID and creation time
- File upload health monitoring now provides additional diagnostic logging (#57133)
- Data source processing idle memory usage has been reduced by approximately 60%, from around 1.5 GB to 650 MB, by consolidating background workers and loading resource-intensive libraries only when required
- Data source processing now automatically derives per-process memory limits from the container memory limit
- Ingestion processing has been made more robust
Fixes
Datasets & Ingestion
- Missing links between receipts and submissions are now created correctly
- Aborted submissions now display the reprocess action correctly, without an unnecessary API call when viewing actions (#58034)
- Reprocessing no longer fails when initiated from affected dataset submissions
- Partial mapping token handling has been corrected (#57374)
- Error processing for non-dedicated datasets now records dataset logs correctly, handles bulk creation failures, retries where appropriate, and waits for asynchronous operations to complete
- Failed ingestion statistics are now saved correctly in ingestion reports
- Quarantine actions can now be processed in batches
Data Management
- Edge editing now works correctly (#58148)
- Expensive deletion eligibility checks have been optimized (#57693)
- RabbitMQ consumer health checks no longer generate excessive reset events or user notifications
- Exclusive workers now start correctly
CluedIn.UI.Gql
Features
- Feature flags have been reviewed and aligned for more consistent behaviour across the platform (#58492)
CluedIn.UI
Features
Rules & Survivorship
- Survivorship rules can now select a data part based on a condition (#57785)
- Vocabulary keys now include an
Insignificant value change option, allowing value changes to be excluded from significant record changes (#58124)
- Conflict resolution now highlights vocabulary keys used by rules and collapses hidden keys for easier review (#57438)
Enrichers
- Enricher preview records can now be partially deleted (#56321)
- An advanced cron editor provides greater flexibility when scheduling enricher runs (#58158)
- Alerts can now be displayed beneath enricher controls (#56853)
- Enrichments can now be tested to evaluate their impact on golden records before being applied (#55137)
- A new Enricher Matches screen provides a consolidated view for reviewing multiple enrichment results (#57217)
Golden Records
- Golden Record layouts can now be searched by name, with matching text highlighted, and include improvements to favorite property selection, configuration visibility, unsaved-change warnings, layout refresh, and rendering performance (#57827)
- Golden Record search column customization now supports searching, drag-and-drop reordering, quick move actions, clearer column limits, and protection for fixed columns (#57816)
- Golden records now show which deduplication groups they belong to (#57627)
Manual Data Entry
- Manual Data Entry provides a cleaner form layout, improved record navigation, searchable existing-value selection with value counts, and more predictable select-field behaviour (#57818)
Security & Authentication
- Password entry now shows whether the entered password meets the configured requirements (#57464)
Workflows
- Workflows can now be deleted directly from the workflow list page (#58196)
Data Management
- Topology and Global Data Model capabilities are now available (#58274)
- Configuration can now be imported and exported through the UI to support configuration migration (#58246)
- Organizations can now be reset to their default state or deleted (#52918)
- Dataset Quarantine and Approval tabs now support batch approval and rejection of clues
Notifications
- The redesigned Notification Center updates existing notifications instead of creating duplicates and supports multi-step progress, attachments, single-use actions, filtering, search, and grouping of related notifications (#58012)
Jobs & Scheduling
- Job monitoring provides improved visibility into job execution and progress (#58122)
- New scheduling pages allow jobs and their schedules to be managed through the UI (#58436)
- High-resolution metric retention can now be configured through the “Trim high time resolution metrics” job (#58573)
- Stream history retention can now be configured through the “Stream log retention” job (#58573)
- Stream idle thresholds can now be configured through the “Stream idle check” job (#58573)
- System jobs now clearly indicate when they cannot be paused, rescheduled, or edited (#58573)
AI & Copilot
- AI Agent and AI Job pages now provide chat capabilities and live job updates (#57662)
- AI endpoint cache duration can now be configured (#58440)
Fixes
Mapping & Properties
- The Add Property panel now shows only configured vocabulary keys for constrained lookup keys (#57942)
- Lookup values now work correctly when adding or editing properties (#58139)
- Properties and Relations pages now load more reliably, with improved Relations performance and faster Golden Record searches (#58152)
- Property filtering, vocabulary key selection, preview uploads, and entry of previously unused property values now behave consistently (#58168)
- Mapping pages have been optimized, and the Properties page no longer crashes when searching
- Mapping and entity type views now refresh correctly after changes, including entity icons and deduplication rules (#58600)
Search
- Saved searches using advanced filters can now be updated and removed correctly, with clearer action labels (#58143)
- Editing a recent search now updates the existing search instead of creating a duplicate, while nested conditions display the correct AND/OR operators (#58154)
- Deleting an active saved search now clears it from the current search and no longer leaves an empty saved-search list (#58168)
Golden Records
- Golden Record Search now displays the actual
Description value (#58129)
- Spacing between the Limited Details badge and View Identifiers action has been corrected (#58169)
- Golden Record graph legends now display correctly on Mac (#56729)
- The Golden Record layout page is now hidden where it should not be available (#58623)
Relations
- Relation details now load their properties correctly instead of repeatedly requesting them from the server
- Relation panels no longer break when opening a record from a grouped set, with pagination and panel sizing also corrected (#58087)
Rules
- Rule formula editor icons now render correctly
- Rule Builder now validates empty groups correctly (#56872)
Notifications
- Notification loading and read-state handling have been improved, with notifications tracked correctly per user and better responsiveness under load
Authentication & Access Control
- The login button is now enabled as soon as credentials are entered (#57464)
- Password confirmation is only enabled when both passwords match (#58053)
- User roles can now be removed correctly from the user page (#57885)
Jobs & Audit
- Job monitoring no longer truncates information incorrectly (#58507)
- Job audit logs now display the associated rule name (#58491)
- Obsolete job capability fields have been removed (#58673)
- Job counts are now calculated correctly (#58721)
AI & Copilot
- The AI Agent page no longer crashes when opening details for a rule created by an AI Agent
- AI agent summaries no longer require unnecessary full-detail loads, improving efficiency (#58609)
Streams
- Export Target configuration on Streams pages now retrieves vocabulary keys with a single API request (#58323)
Runtime-Environment
Features
- Create tables for storing test enrichment data (#55137)
- Added storage for AI Job chats and upload discovery (#57662)
- Added index to DataSets, DataSetLogs and Submissions table to improve some queries
- Add indexes for edge and its properties tables to improve getDataSetById query
- Added index to DataSets, DataSetLogs and Submissions table to improve some queries
- Optimized DataSetLogs and DataSetEndpointIdToPushes shared tables performance
- New nullable column
LastStreamedAtApprox added to the Streams table. No data migration required; existing rows default to NULL (#57215)
- Adds a unique notification-delivery identity constraint to Notifications table (#58530)
- Remove the RecurringJobData table, which is no longer used now that the recurring jobs it served are system custom jobs (#58573)
Packages
Connectors
| Name |
Version |
| CluedIn.Connector.AmazonS3.500 |
100.0.0 |
| CluedIn.Connector.AzureDataLake.500 |
100.0.0 |
| CluedIn.Connector.AzureDedicatedSqlPool.500 |
100.0.0 |
| CluedIn.Connector.AzureEventHub.500 |
100.0.0 |
| CluedIn.Connector.AzureServiceBus.500 |
100.0.0 |
| CluedIn.Connector.Dataverse.500 |
100.0.0 |
| CluedIn.Connector.Dataverse.V2.500 |
100.0.0 |
| CluedIn.Connector.FabricOpenMirroring |
100.0.0 |
| CluedIn.Connector.Http.500 |
100.0.0 |
| CluedIn.Connector.OneLake.500 |
100.0.0 |
| CluedIn.Connector.SqlServer.500 |
100.0.0 |
| CluedIn.PowerApps.500 |
100.0.0 |
Enrichers
| Name |
Version |
| CluedIn.ExternalSearch.Providers.DuckDuckGo.Provider.500 |
100.0.0 |
| CluedIn.ExternalSearch.Providers.PermId.Provider.500 |
100.0.0 |
| CluedIn.ExternalSearch.Providers.Web.500 |
100.0.0 |
| CluedIn.Provider.ExternalSearch.Bregg.500 |
100.0.0 |
| CluedIn.Provider.ExternalSearch.BvD.500 |
100.0.0 |
| CluedIn.Provider.ExternalSearch.ClearBit.500 |
100.0.0 |
| CluedIn.Provider.ExternalSearch.CompanyHouse.500 |
100.0.0 |
| CluedIn.Provider.ExternalSearch.CVR.500 |
100.0.0 |
| CluedIn.Provider.ExternalSearch.DnB.500 |
100.0.0 |
| CluedIn.Provider.ExternalSearch.Gleif.500 |
100.0.0 |
| CluedIn.Provider.ExternalSearch.GoogleImages.500 |
100.0.0 |
| CluedIn.Provider.ExternalSearch.GoogleMaps.500 |
100.0.0 |
| CluedIn.Provider.ExternalSearch.KnowledgeGraph.500 |
100.0.0 |
| CluedIn.Provider.ExternalSearch.Libpostal.500 |
100.0.0 |
| CluedIn.Provider.ExternalSearch.OpenCorporates.500 |
100.0.0 |
| CluedIn.Provider.ExternalSearch.Providers.VatLayer.500 |
100.0.0 |
| CluedIn.Provider.ExternalSearch.Providers.ArtificialIntelligence.500 |
100.0.0 |
| CluedIn.Provider.ExternalSearch.Providers.AzureOpenAI.500 |
100.0.0 |
| CluedIn.Provider.ExternalSearch.RestApi.500 |
100.0.0 |
Crawlers
| Name |
Version |
| CluedIn.Crawling.MasterDataServices.500 |
100.0.0 |
| CluedIn.Purview.500 |
100.0.0 |
Other
| Name |
Version |
| CluedIn.Vocabularies.CommonDataModel.500 |
100.0.0 |
| CluedIn.EventHub.500 |
100.0.0 |
| CluedIn.EnterpriseFlows.PowerAutomate.500 |
100.0.0 |
Controller
| Docker Image |
Tags |
| cluedin/controller |
2026.03.00, 2026.03, 5.0, 5.0.0, 5.0.0_155592 |
Gql
| Docker Image |
Tags |
| cluedin/cluedin-ui-gql |
2026.03.00, 2026.03, 5.0, 5.0.0, 5.0.0_155589 |
Microservices
| Docker Image |
Tags |
| cluedin/data-source |
2026.03.00, 2026.03, 5.0, 5.0.0, 5.0.0_155583 |
| cluedin/data-source-processing |
2026.03.00, 2026.03, 5.0, 5.0.0, 5.0.0_155583 |
| cluedin/data-source |
2026.03.00, 2026.03, 5.0, 5.0.0, 5.0.0_155583 |
| cluedin/data-source-processing |
2026.03.00, 2026.03, 5.0, 5.0.0, 5.0.0_155583 |
| cluedin/data-source-submitter |
2026.03.00, 2026.03, 5.0, 5.0.0, 5.0.0_155583 |
| cluedin/data-source |
2026.03.00, 2026.03, 5.0, 5.0.0, 5.0.0_155583 |
Runtime
| Docker Image |
Tags |
| cluedin/neo4j |
2026.03.00, 2026.03, 5.0, 5.0.0, 5.0.0_155595 |
| cluedin/openrefine |
2026.03.00, 2026.03, 5.0, 5.0.0, 5.0.0_155595 |
Server
| Docker Image |
Tags |
| cluedin/cluedin-server |
2026.03.00, 2026.03, 5.0, 5.0.2, 5.0.2_155929, 5.0.2_155929-alpine, 5.0.2-alpine, 5.0-alpine |
| cluedin/cluedin-server |
2026.03.00, 2026.03, 5.0.2_155929-ubuntu, 5.0.2-ubuntu, 5.0-ubuntu |
| cluedin/nuget-installer |
2026.03.00, 2026.03, 5.0, 5.0.2, 5.0.2_155929, 5.0.2_155929-alpine, 5.0.2-alpine, 5.0-alpine |
| cluedin/nuget-installer |
2026.03.00, 2026.03, 5.0.2_155929-ubuntu, 5.0.2-ubuntu, 5.0-ubuntu |
Ui
| Docker Image |
Tags |
| cluedin/ui |
2026.03.00, 2026.03, 5.0, 5.0.0, 5.0.0_155586 |