The connector estate
Which source systems DataBridge ingests from, and what each can do.
By the end you’ll be able to
- Read the connector capability matrix and know where it comes from.
- Understand the four headline capabilities: incremental sync, sampling, code-lists and data-dictionary.
- Pick the right adapter for a given source-system / region / auth shape.
DataBridge talks to a UK + AU connector estate that spans the major student information systems. The live list is generated from `apps/api/src/adapter-registry.ts` into `docs/CONNECTOR_MATRIX.md` — that document is the source-of-truth for vendors, regions, transports, auth modes, and per-connector capabilities. Connector ids you will see in the product include `sits-api`, `sits-file`, `banner-ethos`, `banner-oracle`, `workday-raas`, `sjms5`, `callista`, `techone-student`, `peoplesoft-cs`, `colleague-ethos` and `unit4-student`.
Every adapter publishes an `AdapterCapabilities` record — exactly four headline flags plus auth and rate-limit hints. `supportsIncremental` says the adapter can read only what changed since a watermark (efficient nightly syncs). `supportsSampling` says you can pull a small slice to verify reachability and shape without dragging the whole dataset. `supportsCodeLists` says the source publishes its reference data (course codes, fee statuses, ethnicity codes…) so DataBridge can pin its crosswalks. `supportsDictionary` says the source publishes a machine-readable schema, which powers drift detection without manual schema upkeep.
Adapters also declare a `preferredAuth` (`bearer`, `basic`, `oauth2`, `db-credentials` or `file`) and a `rateLimitHintRps` so the scheduler can pace itself politely. These shape what the IT team needs to provision: an OAuth client, a read-only DB user, a service account, or an SFTP drop.
Choosing an adapter is therefore a small decision tree: which vendor (Tribal/Ellucian/Workday/Oracle/Unit4/SJMS/TechnologyOne), which transport (REST vs Oracle JDBC vs Postgres vs SFTP), which auth your security team will sign off on, and whether you need `supportsCodeLists`/`supportsDictionary` for your downstream workflows. The capability matrix in the docs answers all four at a glance.
Reminder: the generic `csv` and `jdbc` ingest paths are documented in `docs/CONNECTORS.md` but are not in the matrix — they're escape hatches, not first-class adapters.
Walkthrough
- Open Adapters
1.Open the Adapters screen
The Adapters screen is the in-product mirror of the connector matrix. Each card shows an adapter id, vendor, region and its four capability flags.
- Compare adapters
2.Compare two adapters
Pick a UK adapter (e.g. `sits-api`) and a Global one (e.g. `workday-raas`) and contrast their auth, transport and capability flags. Notice that not every adapter supports code-lists or a data dictionary.
- Open Integrations overview
3.Cross-check the matrix doc
Open the Integrations overview to see how the live estate is presented in-product, then sanity-check against `docs/CONNECTOR_MATRIX.md` for the regions and auth distribution. Always trust the generator, never a screenshot.
Your turn
Open the Adapters screen, then identify one adapter that supports both code-lists and a data dictionary, and one that supports neither. Note the implications for drift detection and crosswalk pinning.
Hint: Use the 'Open Adapters' step — the capability flags are on each card. Per `docs/CONNECTOR_MATRIX.md`, `sits-api` and `banner-ethos` are the two adapters that ship with both `supportsCodeLists` and `supportsDictionary`.