DFS Lite Connectors

Use connectors to bring data from external sources into FactVerse. A connector stores the source type, connection settings, sync behavior, status, and recent error information.

Prerequisites

Before creating a connector, confirm:

• the source owner has approved read access;
• the network route from DFS or DFS Proxy to the source is available;
• credentials, tokens, certificates, or keys are prepared;
• expected source fields, tags, topics, tables, or files are known;
• sync frequency and data freshness expectations are agreed;
• downstream consumers are identified, such as Inspector, AI Agent, predictive maintenance, or BI.

Connector lifecycle

Create a connector

1. Open Data Integration > Connectors.
2. Select Add Connector.
3. Choose a source type.
4. Choose a template if one fits the source.
5. Enter connection settings.
6. Choose sync strategy.
7. Test the connection.
8. Save the connector.
9. Start it immediately, or leave it disabled until mapping is reviewed.

Choose a source type

Source type	Prepare before configuring
OPC UA	Endpoint URL, security mode, credentials, namespace or node information.
BACnet	Network route, device information, object list, polling expectation.
Modbus	Host, port, unit ID, register map, data types, scale factors.
MQTT	Broker URL, topics, payload format, credentials.
REST	Base URL, authentication method, request path, response field paths.
CSV	File, delimiter, header rules, timestamp field, key columns.
JDBC	Database URL, credentials, table or query, timestamp and key columns.
Fabric	OneLake workspace and service-principal details for file inventory, or SQL endpoint details for tabular rows.

Before choosing actions or a sync strategy, check Connector Types and Capabilities. Capabilities differ by connector. In particular, MQTT is inbound subscribe-only, CSV is file import rather than a live connection, and OneLake mode returns file references rather than decoded rows.

Source data inputs

Capture these inputs before the first sync:

Input	Why it matters
Source owner	Needed for access approval and value interpretation.
Source path list	Defines which tags, fields, topics, registers, columns, or files should be inspected.
Time field	Required for time-series quality checks and downstream analysis.
Asset or equipment key	Needed to map source values to operational targets.
Unit and scale	Needed for range checks and conversion.
Expected update interval	Used to detect stale data and sync drift.
Data classification	Confirms whether the data can move into the target tenant or environment.

Name connectors clearly

Use names that help operators understand the source without opening it.

Good connector names include:

• site or plant;
• source system;
• data domain;
• line, area, or equipment group.

Examples:

• Singapore DC BMS meters
• Plant A Line 3 OPC UA
• Warehouse AGV MQTT status
• CMMS work order REST feed

Set sync strategy

Strategy	Use when	Operator check
Realtime	The matrix marks subscribe as supported.	Confirm events arrive continuously.
Interval	The matrix marks poll read as supported.	Confirm interval and expected latency.
On demand	The matrix marks on-demand sync as supported.	Confirm the user or job that triggers sync.

Test the connector

Run Test Connection before saving or starting.

A passing test has connector-specific meaning. For example, Modbus tests only the TCP socket, BACnet can pass after discovering a different device, MQTT does not validate topic traffic, and CSV confirms only that parsed file data is currently loaded. Continue with the supported data-path checks before production use.

If the test fails:

Symptom	Check
Authentication error	Credentials, tokens, certificates, source-side permissions.
Network timeout	Host, port, VPN, firewall, allowlist, proxy route.
Protocol error	Security mode, topic, register map, endpoint path, payload format.
Empty result	Query, topic, node selection, file content, source-side filter.

Start, pause, resume, and sync

Connector List and Connector Detail support connector operations.

Action	Use
Start	Begin data collection for an enabled connector.
Pause	Stop collection while preserving configuration and historical data.
Resume	Restart a paused connector.
Sync	Trigger on-demand sync only when the capability matrix marks it supported.
Delete	Remove the connector configuration after confirming it is no longer needed.

Batch actions can start, pause, or delete multiple selected connectors. Use batch delete carefully because downstream mappings and operational expectations may still refer to the connector.

Expected output

A usable connector should produce:

• a saved connector with clear source name and owner context;
• a passing connection test;
• a supported browse, poll-read, subscription, preview, or file-import result that matches the intended source;
• a selected sync strategy;
• visible sync history after the first run;
• enough source metadata for mapping and quality review.

Monitor connector detail

After the connector starts, review:

• connector status;
• last sync time;
• throughput;
• data point count;
• quality score;
• recent errors;
• mapping status;
• sync logs.

Validation checklist

A connector is ready for mapping and broader use when:

• connection test succeeds;
• the connector name identifies the source;
• source owner has confirmed the source purpose;
• the connector's supported validation path shows expected data;
• sync strategy matches the operational need;
• recent errors are clear;
• source credentials and network route are approved for the project.

Next step

After a connector is reachable, continue with Mapping Source Fields.