Skip to main content

Release Notes - Datadios v1.9.0

Release Date: 2026-02-11 Release Type: Minor Released By: DataDios Team

What's New​

🚀 New Features​

Dataquality Dashboard and AI Rule Generation​

Added Dataquality Dashboard to monitor data quality metrics and identify data quality issues.

AI Studio with LLM​

Added AI Studio with LLM support to generate queries; users just need to provide a prompt, and it will generate the queries for them.

Agents (First Release)​

Introduced Agent Workflows with the Code Analysis workflow:

  • Code Structure Analysis to map dependencies and surface issues
  • Complexity Analysis to classify functions and highlight risk areas
  • Human-in-the-loop approvals between steps
  • Optional RAG datasource support for additional context

Supported Types Page​

Added Agents → Supported Types to view available agent workflows and their steps.

🔧 Bug Fixes​

Fixed user deletion issues. You can now delete users and their related data reliably.

Deleting a user now also removes their sharing entries, so recipients no longer keep listings that can't be opened:

  • Resources the deleted user owned are removed from every recipient's shared list, including resources shared with all users.
  • Grants the deleted user had received are revoked for that user alone — other recipients of the same shared resource keep their access.
  • Deleting a user is permanent and is not reversible; entries orphaned by deletes made before this release are not cleaned up retroactively.

See Resource Center - Delete User Accounts for details.

SQL support for sheared datasources​

Added SQL support for sheared datasources. shared datasources can be shown in the SQL editor. and you can perform queries on those

Scheduler workflow for scheduled query diff jobs​

Fixed Schedueler reletes issues. Now you can run the scheduled query diff jobs and get the results.

Coverage: which of your data has no rules at all​

The Quality Overview could tell you how your rules were doing, but not how much of your data had any rules. A new Coverage card answers that — of the objects the platform has listed, how many carry at least one rule, plus a table of the ones that do not. It appears on the fleet Overview and on each hub's page, narrowed to that hub.

It is deliberate about when it will not answer. Coverage matches listed objects against the objects rules say they target; when enough of those targets cannot be matched to move the figure by more than a point, the card shows how many failed to match and withholds the percentage rather than reporting a number a mismatch dragged down. It also says when nothing has been listed yet, instead of reporting zero coverage for a scope whose workflows have not run.

A page for each monitored object​

Clicking an object on a quality hub used to open a flat list of the rules targeting it. It now opens that object's own page with three tabs — Overview (its stat cards and run-status trend), Rules (the same list as before), and a new Columns tab showing which of the object's columns have rules and how those rules are running. The row you click also names the object's data source, so two sources holding a table of the same name open different pages.

Data Quality counts objects by their full name​

The Data Quality dashboard identified a monitored object by its table name alone, so two data sources under one hub that each held a table of the same name were counted as one object and their run results were averaged together — a table failing in one source could be reported as healthy. Objects are now identified by their data source and fully qualified name. Monitored Objects can go up after this release, and an object that was previously one row may now appear as two, one per source. Rules that target a model rather than a table (AI evals) are no longer counted as objects.

Snowflake workload extraction cadence​

Workload extraction for Snowflake now checks in hourly instead of every 15 minutes. ACCOUNT_USAGE publishes 45 minutes to 3 hours behind, so the faster cadence was waking your warehouse four times an hour to ask for data that did not exist yet. Nothing you collect changes — only how often DataDios asks for it.

Two metric groups, SF Daily Compute Usage and SF Warehouse Load Usage, have also been removed. Nothing in the product displayed them, and each paid for a full extraction cycle on every check-in. They disappear from the metric-group picker, their scheduled jobs stop, and their views are dropped on the next sync. Rows already collected are retained.

If you had customised the extraction schedule for this data source, re-check it after upgrading. See Snowflake Permissions for the grants extraction needs.

Upgrade Instructions​

Deployment​

Pull the latest Docker image

Post-deployment​

  • Refresh the browser after upgrade
  • Verify LLM and input datasources before running Agent workflows

Support​

For technical questions or issues, please contact our support team.

Released by Team Datadios | Document version: 1.0