Connectors
Connectors link ArchNGN to external systems, allowing you to ingest metadata and keep your architectural models aligned with what is actually deployed. You will find Connectors in the left sidebar.
For the per-system setup guides, see the Connector Setup section, starting with GitHub.
Available Connectors
Section titled “Available Connectors”| Connector | Source Kind | What It Ingests |
|---|---|---|
| GitHub | Code | A Git repository. The AI explores the codebase and maps it to Application, Technology, and Data objects |
| Amazon S3 | Documents | Architecture documents, IaC, and configuration files from an S3 or S3-compatible bucket |
| Azure Blob Storage | Documents | The same, from an Azure Blob Storage container |
| Google Cloud Storage | Documents | The same, from a GCS bucket |
| SharePoint | Documents | Publicly shared SharePoint/OneDrive files (“Anyone with the link”) |
| AWS Cloud | Cloud infrastructure | The as-built resource inventory of an AWS account, read-only |
| Azure Cloud | Cloud infrastructure | The as-built resource inventory of an Azure subscription, read-only |
| GCP Cloud | Cloud infrastructure | The as-built resource inventory of a GCP project, read-only |
Confluence is on the roadmap and appears greyed out in the connector picker.
The three source kinds produce different things:
- Code sources map the repository to metamodel objects and associations, the as-detailed view of your applications.
- Document sources are documentation about your architecture, not the deployed architecture itself, so they map to Reference Library registry items and Enterprise Glossary terms rather than creating architecture objects.
- Cloud infrastructure sources map your live cloud estate to Technology and Data objects (the as-built view) and infer the relationships between them (routing, messaging, data access, network membership).
ArchNGN automatically links the layers: after runs complete, a code-derived Application object is connected to the cloud-derived Technology object that runs it (a Deployed To association), matched on deterministic evidence such as container image references and resource tags. Ambiguous matches are held as suggestions for an architect to confirm rather than linked automatically.
Adding a Connector
Section titled “Adding a Connector”- Navigate to Connectors in the left sidebar.
- Click New Connector.
- Select the connector type.
- Fill in the connection details and credentials. See the per-connector setup guides for the prerequisites each type needs.
- Optionally link the connector to a Business object (Business Ownership). Valid associations to it are created automatically after each run.
- Save the connector.
Credential Handling
Section titled “Credential Handling”Secrets you enter (tokens, access keys, client secrets, service-account keys) are moved into an encrypted secrets store on save. They are never persisted in the connector configuration and never shown again in the UI; the form shows a “securely set” placeholder instead. To rotate a credential, open the connector and type or paste the new value over the placeholder.
Analysis Mode (Document Sources)
Section titled “Analysis Mode (Document Sources)”Cloud storage and SharePoint connectors have an Analysis Mode:
- Formal artefacts only (default): only documents that formally define governance (a policy, standard, or decision with a defining title or classification) become Reference Library registry items. Other files are skipped.
- One item per document: catalogues every file as a registry item unless it clearly isn’t governance content.
Either way, every file analysed and the reason it was kept or skipped is recorded in the run summary.
Connector State
Section titled “Connector State”Each connector reports its current state:
| State | Meaning |
|---|---|
| Connected | Authenticated and ready to sync |
| Provisioning | Being set up or running its first sync |
| Disabled | Paused; no syncs run until re-enabled |
| Issue | A run failed or credentials need attention |
Running a Connector
Section titled “Running a Connector”From the Connectors page:
- Locate the connector you want to run.
- Click Run to start the sync.
- The connector runs as a background job. Monitor progress on the Jobs page.
Only one run per connector is active at a time. Clicking Run while a sync is in progress attaches you to the running job instead of starting a duplicate. Different connectors run in parallel.
Each run creates or updates objects and associations in the active workspace based on the external data. Objects are classified against your account’s effective Metamodel, and the connector links them to the Reference Library guidance that governs them where it applies. Run results and history are stored for each connector.
Re-runs are stable: ArchNGN matches incoming resources to existing objects by their stable identity (for cloud connectors, the resource ID or ARN) before falling back to name matching, so a re-run updates objects in place rather than deleting and recreating them.
Managing Connector Data
Section titled “Managing Connector Data”- View history: See all previous runs, their status, and results
- Re-run: Trigger a fresh sync at any time
- Delete: Remove a connector, with the option to also delete all objects and associations it created
A connector only owns the associations it created. Links drawn by other connectors or by cross-layer resolution are never removed by another connector’s run.
Token Usage
Section titled “Token Usage”Connector sync jobs consume AI tokens from your account allowance. Token costs vary depending on the AI mode selected. For a full breakdown of costs by job type and AI mode, see Usage.