DataLug API
The DataLug API connector calls an API your account built in API Studio. The connection names the API, the version and the environment. The actions come from the version's published spec, so paths, parameters and columns are filled in for you.
When to use it
- A pipeline reads records from an API your account deployed.
- A pipeline writes records to one of your own APIs.
- A Lookup node adds fields from one of your own APIs to each row.
For an API outside your account, use the Generic REST API connector.
What you need
- The API Studio add-on. Without it the connector is shown locked.
- An API with a published version, deployed to at least one environment. See Publishing and versions and Deploying an API.
No key or password is entered on the connection.
Connect
| Field | What to enter |
|---|---|
| API | An API with a published version. |
| Version | Latest promoted, or one published version. |
| Environment | An environment where the API is deployed. |
Latest promoted follows the version deployed in the chosen environment, so a promotion or a rollback is picked up on the next run. A pinned version must be the one deployed in that environment, or the run stops with an error.
Calls go through the same address, validation, policies and call log as any other client. DataLug creates a client named DataLug pipelines for each environment the first time a pipeline calls it, and holds its key. That client is listed on the Clients page and cannot be revoked there.
Stored secrets are never sent back to the browser.
Supported operations
| Action | Node | Comes from |
|---|---|---|
| Read | Source | GET operations. |
| Look up | Lookup | GET operations keyed by a path parameter or a required query parameter. |
| Write | Load | POST, PUT, PATCH and DELETE operations. |
Choose an Operation in the action panel. The panel then shows its path and
query parameters as Parameters. A value can be fixed text or a row field,
such as {field:customerId}.
The columns come from the operation's response schema on a Source and from its request body schema on a Load, so Map Fields needs no preview. Read-only fields are left out of a Write unless the schema requires them.
Pagination follows the API's style: offset and limit, page number, cursor token
or the Link header. Rows are read from the page's array, such as items.
Operations that return JSON Lines, server-sent events or files, and HEAD and OPTIONS operations, are not offered.
Limits
Pipeline calls are not counted against your included call allowance. They appear in the call log marked as pipeline calls.
The API's policies still apply, including rate limits and quotas. A pipeline
that goes past a limit gets the same 429 response as any other client.
Troubleshooting
- The API is not in the list. Only APIs with a published version are listed. Publish a version first.
- No environments to choose. The API is not deployed. Deploy the version to an environment.
- The run stops with "no resolved edge target". The API is not deployed in the environment, or the pinned version is not the one deployed there. Choose Latest promoted or deploy the version.
- A Write returns 400. The edge validates each request against the spec. Map every required field of the operation.
- No operations to choose. The version has no operation the connector supports for this node. Check the method and response type in the spec.
Related
- API Studio and Clients and keys.
- Source node, Load node and Lookup node.
- Generic REST API for APIs outside your account.