Data Observability platform Help

Objects

Objects page gives users an overview of all database objects available, allowing to create a data catalog to easily manage data assets.

The Objects view can be accessed on the Catalog page under the Objects tab.

catalog_objects.png

Each catalog object consists of:

  • Data quality metric - calculated based on test cases and profilings

  • Name - object name in database

  • Attributes - number of columns for object

  • Row count - latest known row count from profiling

  • Test cases - related test cases (click to navigate to Test cases page)

  • Profiling rules - related profilings (click to navigate to Profiling page)

  • Description - business definition (if exists) or database comment

Object detailed view

Clicking on an object name opens a detailed view, allowing users to get insights on attribute level. Users can view metadata and metrics, add comments, and manage related business terms from Glossary.

catalog_object_detailed.png
  • Audit history - catalog_object_audit_history Browse this object's history via Audit history.

  • Alerting - catalog_object_entity_alerting Configure alerting for object changes via Entity alerting.

  • Database comment - technical comment from database metadata

  • Dataset - set the dataset the object belongs to

  • Business definition - custom free text field to describe the object

  • Labels - add custom labels to group objects

  • Custom fields - list of custom fields enabled for the entity type Object.

The object detailed view contains five tabs:

  • Attributes - attributes of the object

  • Lineage - opens the objects Lineage

  • Dependencies - shows all upstream and downstream dependencies of this object across the data landscape

  • Test cases - shows the test cases related to the object. You can also create test cases directly from this tab.

  • Profilings - opens the Profiling page for this object

Clicking on a row in the attributes table will open that attribute's detailed view.

Dependencies

The Dependencies tab shows all upstream and downstream column-level dependencies for the catalog object, giving a full picture of where this object is used and what it depends on.

catalog_object_impact.png

The tab displays a filterable table where each row represents a column-level dependency:

  • Direction - Downstream means this object feeds data into the related object; Upstream means this object consumes data from it

  • Related object - connection, schema, and object of the related catalog entity

  • Related column - the source and target column pair that form the relation. Arrow color indicates the relation type:

    • Green / Blue - directed relation (downstream / upstream)

    • Gray - foreign key relation

    • Orange - general related relation

  • Quality - data quality bar showing the pass/fail/error/not-executed test case result ratio for the related column

  • Lineage - button showing the number of steps in the relation path; clicking it opens the column-level lineage view

The same dependency view is also accessible directly from a test case via the impact and root cause analysis.

Creating test cases from object view

From the Test cases tab in the object detailed view, you can create test cases directly for the current object using standard validations.

  1. Open an object's detailed view and navigate to the Test cases tab

  2. Click the "New test case" button

  3. Select a predefined validation type from the modal

  4. The object context (connection, schema, object) is pre-filled automatically

  5. Fill in remaining details and click "Create" or "Create and Execute"

After creating a test case, the object's data quality metric and test case count refresh automatically to reflect the new test.

The Objects tab has a smart search box that suggests scoped filters as you type. Type any text and within ~500 ms a dropdown appears with the matching suggestions, grouped by type. Pick a suggestion to apply it as a filter to the objects list, or press Enter to accept the top result.

catalog_search.png

The following suggestion types are available:

  • Text - match the typed string against object and column names and descriptions (free-text search)

  • Object name - match a specific catalog object (table / view name)

  • Column name - match a specific column across all objects

  • Object label - match a custom label assigned to objects

  • Term - match a business glossary term linked to a column; the suggestion shows the term's color dot

Suggestions respect any sidebar filters already applied (datasets, connections, etc.). When that's the case, the dropdown shows a small "filters applied" notice so you know the suggestion list is narrowed to the current scope. Click the × button on the right of the search input to clear the active suggestion and return to the unfiltered list.

Description enrichment with AI

Catalog objects and columns can be enriched with AI-generated business descriptions in bulk from the Objects tab.

  1. Open the Objects tab and click the "Enrich" action (star icon) in the toolbar

  2. Select step - choose a connection and schema, then check the objects to enrich. For each object, expand its row to also pick individual columns. Optional toggles:

    • Enrich objects - generate descriptions for the object itself

    • Enrich attributes - generate descriptions for the selected columns

    • Include column values - let the AI see sample column values to improve accuracy (off by default for sensitive data)

    • Use existing glossary terms - bias the AI toward existing business terms instead of inventing new wording

    • Domain context - free-text instructions added to every prompt (e.g. "this is a retail orders system, prefer terms like SKU and order line")

  3. Preview step - review the AI-generated descriptions for every selected object and column. Edit any suggestion inline before applying.

  4. Click "Apply" to write the approved descriptions to the catalog. Skipped or failed entries are reported in an alert at the end of the run.

Existing descriptions are overwritten when enrichment is applied, so review the preview carefully before confirming.

Datasets

Datasets can be created to group specific catalog objects. Datasets are found on the sidebar. New datasets can be created by clicking on Plus icon. Clicking on a dataset will filter Catalog view to display objects in selected dataset.

catalog_datasets.png

Datasets also support configuring custom fields.

catalog_dataset_edit.png

Assigning objects to datasets

Objects can be added to datasets by dragging the Dataset icon to any of the datasets. To remove object from any of the datasets, drag it to "All objects" dataset.

catalog_object_add_directory.png

Custom metadata

Clicking the "New Object" button Metadata new object button in the left menu opens a modal to create a new custom object.

newObjectModal
  • Connection – Select the connection under which the new metadata will be added.

  • Schema – Click the schema field to view suggested schemas based on the selected connection, or enter a new schema manually.

  • Object – Enter a name for the new object.

  • New Column button – Adds a new row for inserting a column.

  • Manage columns

    • Name – The name of the new column.

    • Data type – Click to select an existing data type or enter a new one manually.

    • Remove column – Click the trash icon Metadata trash icon to remove a column.

  • Save – Confirms and saves all inserted values.

Modifying existing metadata can be done from the object detail view by clicking the "Manage Columns" button Metadata edit button. This opens a modal for editing existing columns.

editModal
  • Connection – The objects connection will be preselected.

  • Schema – The objects schema will be preselected.

  • Object – The object name will be preselected.

  • New Column button – Adds a new row for inserting a column.

  • Manage columns

    • Name – The name of the column.

    • Data type – Click to select an existing data type or enter a new one manually.

    • Remove column – Click the trash icon Metadata trash icon to remove a column.

  • Save – Saves all changes.

  • Delete objectUse with caution. This will remove the object and all its columns.

22 May 2026