Help & Documentation
Open Support Ticket

Best Practices & First Principles: Collibra

This article contains a general overview of how Collibra can be connected to Metric Insights, what specialists will be involved in establishing that connection and what choices have to be made before performing the connection.

Relationship Between Metric Insights and Collibra

The recommended architectural model when it comes to Collibra, is the "Source of Truth" model. This means, that to maintain data integrity, Collibra functions as the single "Source of Truth," while Metric Insights (MI) acts only as the consumption layer.

Pros:

  • Automated data collection and updating without manual interference from Users.

Cons:

  • The process of collecting data has to be properly set up, so no manual corrections are necessary.

Metadata and Glossary Synchronization Process

Glossary Terms and Metadata fields can be created and populated manually in Metric Insight, but according to "Source of Truth" model it has to be done automatically. It is also more convenient for Users that way. However, to automatically and correctly tag Metric Insights Elements (Reports, for example) with Collibra Glossary Terms and Metadata items, the source Dataset must contain a common identifier that exists in both systems. For example, a Term named "Revenue" in Collibra needs to know it belongs to the "Q4 Sales Dashboard" in Metric Insights.

This problem is solved by relying on a specific column in the Dataset to bridge the gap. You must configure the Dataset to include one of the following keys:

  •  External BI Tool Report ID: (Recommended) The unique GUID from Collibra database. This is the most robust method as it persists even if the Report name changes.
  •  Element Name: The exact name of the Report in Metric Insights. This option is more brittle; if a Report is renamed, the association link may break.
  •  External GUID: A specific external global identifier if maintained across systems.

The process of connecting Collibra to Metric Insights with all the steps to also synchronize Glossary Terms and metadata is described in Collibra Metadata Sync & Enrichment Workflow Playbook.

Required Administrator Roles

RoleResponsibilities
Collibra Administrator
  • Generate Client ID and Secret for API access.
  • Ensure the Collibra Dataset includes the relationship between the Business Term and the Technical Asset (the BI Report). This may require setting swap_relation_target_source to "true" if relations are stored as Term-to-Asset.
Identity (IAM) Lead Ensure "Owner" names in Collibra exactly match User IDs or Identity Profiles in MI to prevent assignment failures.
Metric Insights Administrator
  • Create Data Source for Collibra and Dataset to collect data from Collibra.
  • Configure the Glossary Autoimport and Glossary Auto Synchronization.
  • Pre-define Custom Fields prior to syncing and synchronize them with Collibra Metadata.

Prerequisites

Before technical implementation, make sure to:

  • Ensure that the combination of Glossary Term and Section is unique within the source data. If duplicates exist, the system will skip those records to prevent conflict.
  • Identify which Metadata attributes are "Standard" (like Description) versus "Custom" (like Formula, or URL Reference). MI requires Custom Fields to be created in the Admin console before data can be synced into them.
  • When planning Custom Fields, strictly define the input type (e.g., "Text for formulas", "URL" or "Text" for links) to ensure the Custom Field Editor accepts the incoming data format.
  • Determine if the synchronization should be a Snapshot (assuming that during the update only new terms are loaded) or a full reload. For maintaining a synchronized mirror of Collibra, a full load strategy is generally recommended to capture all updates to existing definitions.