Import data from FTP

FTP pipelines turn remote delimited files into instrument readings. They combine a granted external connection, a remote target, instrument metadata, and a reusable parser.

Before you start

An organization administrator must configure an FTP external connection and grant your project access to an allowed root folder. The connection settings support automatic, explicit FTPS, implicit FTPS, or plain FTP modes. Use the configured mode appropriate for your upstream server and review its health.

Your project role must allow pipeline authoring and access to its parser and source. The FTP Browser lets authorized users browse granted roots and download source files without modifying the remote server.

Create the pipeline

Open Ingestion → Pipelines, add an FTP pipeline, and complete the wizard:

1. FTP target: select the granted connection and browse to the folder or file.
2. Choose New files in a folder or New rows in one file. For a folder, configure a filename pattern if needed.
3. Instrument metadata: set the base instrument name, destination site, and pipeline timezone. Review the Override Data Points choice for duplicate readings.
4. Parser and preview: create a parser, select a saved parser, or use AI generation if enabled on the deployment.
5. Run the preview, review its mappings, data, and diagnostics, then save the pipeline.

The base instrument name is also the pipeline’s display name. With a parser that detects instrument names, the final name follows the base-and-detected naming rule shown in the preview. Check it before saving; instrument names are fixed after creation.

Review the preview

Preview shows at most ten successful parsed points across the selected input. Inspect the final instrument names, timestamp interpretation, variable names, converted values, and value types. Review diagnostics even when some rows succeed.

A timezone mistake can shift an otherwise valid reading into the wrong period. Compare a source row with its preview timestamp. See FTP parsers for layouts and conversions.

AI generation proposes parser configuration. Review and preview the result using representative files before using it.

Operate and troubleshoot

The pipeline detail separates Enabled, Computed health, and Lifecycle. Use Activity for execution outcomes and Files & attempts for file revisions, valid and rejected counts, and processing details.

Reprocess current file targets one currently listed remote file revision, reads the full file, and does not move the normal incremental cursor. Review the override policy and diagnostic cause before reprocessing.

Archiving stops future polling, although started work may finish. It preserves the stored enablement value for restoration. Check instrument values and runs after commissioning or changing a parser.