Skip to main content

Data Observability Overview

The Overview sub-tab on a Table’s Data Observability page gives you an instant snapshot of that table’s quality health — without digging through individual test cases. Data Observability Overview tab showing Test Run Summary, Test Coverage by Dimension, and Dimension-wise Test Results

Test Run Summary

The top section shows four key metrics for the table:
  • Failed: Number of tests that failed in the selected time window, with a trend chart showing how failures have changed over time
  • Success: Number of tests that passed, with a trend chart
  • Open Incidents: Active incidents triggered by test failures — click View Open Incidents to jump straight to the incident list
  • Upstream Failures: Test failures in tables upstream in your lineage that may be causing issues here — click View Upstream Failures to investigate

Test Coverage by Dimension

The bar chart on the right shows how your tests are distributed across quality dimensions — Integrity, Completeness, Consistency, and so on. Use this to spot gaps: if a dimension has no bar, you have no tests covering that category for this table.

Dimension-wise Test Results

Scroll down to see test results broken down by dimension. Each tab shows the test cases assigned to that category and their current pass/fail status.
  • Integrity: Tests that validate the structural rules of your data — for example, checking that column counts, row counts, or foreign key relationships are within expected bounds
  • Completeness: Tests that check for missing or null values — ensuring required fields are populated where they should be
  • Consistency: Tests that verify data stays stable and predictable across processing steps — for example, comparing values between tables or checking that a column’s value distribution hasn’t shifted unexpectedly
  • Accuracy: Tests that confirm values reflect real-world expectations — for example, that a price column stays within a valid range
  • Uniqueness: Tests that catch duplicate values where only unique entries are expected
  • Validity: Tests that enforce formatting or domain rules — for example, that an email column matches a valid pattern
Dimension-wise Test Results showing tabs for Integrity, Completeness, and Consistency Only dimensions with at least one test assigned to this table appear as tabs. Tabs with failures are the ones to investigate first. This view pairs with the Quality by Dimension page, which lists all available test cases and their assigned dimensions.