Development of a custom endpoint includes minimal coding. The implementation includes:

• Configuration to support user interaction with the custom endpoint.
• The pipelines AutoSync needs to retrieve, load, and synchronize data.

User interaction

When creating a data pipeline, AutoSync users select up to five sources and one target endpoint. For each source endpoint, they specify:

• Credentials and connection information.

• The source schema, if applicable, and the tables, objects, or files to synchronize.

For each target endpoint, they specify:

• Credentials and connection information.
• The target schema.

For each data pipeline, they specify:

• The load type.
• The synchronization schedule. Your endpoint doesn't need to handle this. AutoSync takes care of it.

To expose your custom endpoint, you supply a schema.json file that configures the UI components for credential and load type selection and validation. You also provide a logo in SVG format. Project folder structure shows the location for the schema file and logo. Endpoint schema describes the structure of the schema file.

Pipelines to retrieve and load data

AutoSync hides the complexity of the implementation from users. Under the covers, an AutoSync data pipeline uses pipelines composed of Snaps. You create these pipelines for a custom endpoint. AutoSync handles change capture and reconciliation for you.

Each endpoint requires four pipelines:

• Source: actions, definition, extract, and load
• Target: actions, definition, extract, and load

To create the required pipelines, you need to know the data types the endpoint supports, including precision and scale constraints. In the definition pipeline, you map endpoint types to common data model (CDM) types. Refer to Common data model for details about the common model schema and the supported common model data types. Use the SimpleGoogleSheets and SimpleSnowflake pipelines as examples.

To get started, download and install the SDK using the Maven archetype.