Collibra Integration

What it is

The Collibra integration builds a Collibra-shaped picture of your Tessallite semantic model — your metrics, dimensions, KPIs, glossary terms, and downstream reports, complete with ownership, classifications, and relationships — so you can preview exactly what would land in Collibra, the enterprise data governance and catalog platform.

In this build the integration is a preview / dry-run tool. It reads your model and computes the Collibra payload locally; it does not send anything to Collibra. Live push (writing the assets into your Collibra instance) is not yet available — see the "Not available yet: live connector" section below.

Who it is for

• Modellers who want their models governed in the enterprise data catalog
• Tenant Admins who configure the connection to the Collibra instance
• Data Stewards who manage ownership, classifications, and governance workflows in Collibra

What gets exported

When you preview or dry-run a model export, Tessallite maps these objects into governed Collibra assets with attributes, relations, and responsibilities (computed locally — nothing is written to Collibra yet):

Each asset carries its Tessallite attributes (descriptions, types, formulas, visibility). Relationships link assets together. Ownership becomes Collibra responsibilities for objects that carry an explicit owner — KPIs and downstream assets. Measures and dimensions have no owner field, so they export without a responsibility.

Lifecycle statuses are mapped automatically: deployed → Accepted, draft → Candidate, hidden → Candidate, deprecated → Deprecated.

Key difference from Solidatus

Solidatus is a lineage graph — the main deliverable is a visual map of nodes and edges. Collibra is a governance operating system — the main deliverable is a governed catalog with assets, attributes, relations, responsibilities, statuses, and classifications. The Collibra integration exports richer governance metadata: ownership, stewardship, classifications, and lifecycle states.

Step-by-step: Set up and sync

Step 1 — Get your Collibra details

Before you start, ask your Collibra administrator for:

• Base URL — the web address of your Collibra instance (e.g., https://your-company.collibra.com)
• API Token — a service account token with import permissions
• Community ID — the Collibra Community where assets will be placed
• Domain ID — the Collibra Domain within that Community

Step 2 — Open Model Settings

1. In Tessallite, open the project and model you want to sync.
2. Click the gear icon (⚙) in the top-right corner of the Model Builder.
3. The Model Configuration drawer opens.

Step 3 — Navigate to the Collibra tab

1. In the settings drawer, scroll the tabs until you see Collibra.
2. Click the Collibra tab.

Step 4 — Add a connection

1. Click the Add Connection button.
2. Fill in the dialog:
   • Display Name — a friendly name like "Collibra Production"
   • Base URL — the URL from Step 1
   • API Token — the token from Step 1 (kept encrypted)
   • Community ID — the Collibra Community ID from Step 1
   • Domain ID — the Collibra Domain ID from Step 1
3. Click Save.

Step 5 — Test the connection

1. Click the Test Connection button.
2. In this build the connection check is simulated — Tessallite does not yet contact your Collibra instance. Your configuration is saved, but it is not verified against a live Collibra API.
3. You will see an informational notice that the validation was simulated. This is not a green pass: a wrong URL or expired token will not be caught until live validation is available.

Step 6 — Preview what will be exported

1. Click the Preview Export button.
2. Tessallite builds the governance graph and shows you:
   • How many assets will be created (metrics, dimensions, tables, etc.)
   • How many relations will be created (relationships between them)
   • How many responsibilities will be assigned (ownership mappings)
3. The preview breaks down counts by asset type so you can verify coverage.

Step 7 — Dry run

1. Click the Dry Run button.
2. Tessallite runs the full sync pipeline — builds the graph, maps to Collibra assets/relations/responsibilities, calculates what would change — but does NOT push to Collibra.

Step 8 — Live push (not yet available)

Live push to Collibra is not implemented in this build, so the Sync button is disabled. The dry run in Step 7 lets you confirm exactly what would be created. When a live Collibra client is wired in, this button will push the previewed assets, relations, and responsibilities and record the result in run history.

Available now: preview and dry run

Everything here works today. It reads your model and computes the governance payload locally — nothing is sent to Collibra.

1. Build governance graph — reads all model objects from the snapshot.
2. Map to Collibra format — converts to assets (with attributes + status), relations, and responsibilities.
3. Hash every object — SHA256 fingerprint, ready to drive an incremental diff once live push exists.
4. Record the run — saves the run in history with the exact model snapshot it was built from, the asset/relation counts, and any governance warnings.

The dry run computes what a push would do; it does not contact Collibra and does not create, update, or deprecate anything remotely.

Incremental diff is not available yet. Live push is not implemented in this build, so no remote mapping is ever saved. With no saved baseline to compare against, every dry run reports all objects as new — it cannot yet show which objects changed or stayed the same since a previous sync. When live push lands, the fingerprints above will drive an incremental create/update/deprecate diff.

Not available yet: live connector

Live validation and live push are not implemented in this build. Until a tenant-specific Collibra client is wired in, the following do NOT happen:

• Live push — the Sync button is disabled and the API returns "not implemented". No asset, relation, or responsibility is created or updated in Collibra.
• Remote deprecation — when you remove objects from your model, the dry run records them as removed locally, but nothing is marked Deprecated inside Collibra, because no remote call is made.
• Live connection checks — Test Connection is simulated (see Step 5); a wrong URL, expired token, or missing Community/Domain is not caught until live validation exists.

When the live client lands, this page will be updated with the remote push, deprecation, and troubleshooting behaviour it introduces.

Tips and best practices

• Export the deployed version — Preview and dry run use the deployed model state by default. Draft changes won't appear until deployed (or unless you enable "export draft").
• Ownership flows through — KPIs with an owner and downstream assets with an owner become Collibra responsibilities in the payload. Measures and dimensions have no owner field, so they export without a responsibility.
• Read the governance warnings — A dry run flags objects missing a description or owner, and KPIs / calculated measures whose expression could not be fully resolved.
• Check the recorded snapshot — Each run records the exact model snapshot it was built from.
• Business vs Technical — The export supports separate business and technical views. Hidden columns are included in technical exports.
• Asset type mappings — The default asset type names work with standard Collibra installations. Custom setups can configure the mapping.

Troubleshooting (preview / dry run)

These cover the preview and dry-run surfaces that work today. There is no live troubleshooting yet because no remote call is made.

Related

• Solidatus Integration — The equivalent integration for Solidatus lineage platform
• Model Configuration — All model-level settings
• Business Glossary — Manage business terms
• API Reference — REST API documentation