Pipelines
Pipelines are the core mechanism for moving data in VendorGate. A pipeline connects a source, one or more contracts, and a delivery destination. A single pipeline uses one contract; a workbook pipeline uses one contract per sheet. Pipelines can run on a schedule or on demand.
How a single pipeline moves and validates data
Source
Collect the file
Contract
Validate the file
Delivery
Send it forward
Pipeline components
Every pipeline has at least three parts:
- Source: where the data comes from.
- Contract: the rules the file must satisfy before delivery.
- Delivery: where the data goes after processing.
You can also configure:
- Transformation: to standardize or reshape the file before delivery.
- Execution options: such as fail-fast, which stops processing after a configured number of errors.
Supported sources
VendorGate supports several source types:
- SFTP / FTP: poll a remote folder for new files.
- HTTP API: fetch data from a REST endpoint.
- S3-compatible storage: connect to buckets on Amazon S3, MinIO, or similar services.
- Database: run a query against a supported database.
Supported delivery destinations
After processing, data can be delivered to:
- SFTP / FTP: write to a remote folder.
- HTTP API: post to an endpoint.
- Database: insert into a table.
- S3-compatible storage: write to a bucket on Amazon S3, MinIO, or similar services.
- Lookup dataset: load the output into a reference dataset that standardizers can use.
Creating a pipeline
- Go to Pipelines.
- Click New Pipeline.
- Choose a source type and select the credential that grants access.
- Configure the source details, such as folder path or endpoint URL.
- Select the contract that validates the file. For workbook pipelines, select one contract per sheet.
- Choose a delivery destination and credential.
- Set the schedule or leave it manual.
- Save.
Running a pipeline
You can run a pipeline manually from the pipeline detail page or let the schedule trigger it. Each run produces a log of events, including how many rows were processed and any validation errors.
Pipeline runs
The run history shows every execution, its status, and a link to detailed logs. Use this view to debug failures, monitor throughput, and confirm that files are arriving on time.