How Trinium works

A Trinium ETL pipeline is a set of nodes wired together on a canvas in your browser. Rows stream from node to node, every node runs at once, and the builder shows each node's columns before you run anything.

A Trinium ETL Test run runs a pipeline on a 50-row sample against Staging, writes nothing to any outside system, and shows each node's rows, columns and messages in the builder.

In Trinium ETL you edit a draft, save numbered versions, and deploy a version to Staging or Production. Each environment runs only what was deployed to it, and Production takes a version Staging ran first.

A Trinium ETL Flow runs whole pipelines in order, independently, or depending on how an earlier one ended, and sends a notification when you say. The Flow moves control; pipelines hand data to each other through storage.

Trinium ETL schedules run a pipeline or a Flow on a timer, or run a pipeline when new files arrive in cloud storage or on SFTP. A schedule belongs to one environment and never overlaps itself.

Trinium ETL runtime parameters let a pipeline ask for values when it runs. Type ${name} into any node's settings; the Run dialog, a schedule or a Flow step supplies the value, and built-in date tokens fill themselves in.

A Trinium ETL bookmark remembers what an input has already loaded, so each Live run reads only what is new. A file input remembers the files it read; a database input remembers the largest value of its bookmark column.

A Trinium ETL connection is a saved, encrypted set of settings for one system, with a Staging side and an optional Production side, and an access list deciding who may use it and who may reach Production.

Trinium ETL has eight fixed roles, from User to Super Admin plus a read-only Auditor, and workspaces that share pipelines and Flows between people with View or Edit seats.

Trinium ETL sends email, Slack and Discord notifications when scheduled pipelines and Flows fail, succeed, finish with warnings or are cancelled, to people or to named destinations with their own access lists.

Trinium ETL records every change, and who made it: sign-ins and failed sign-ins, role changes, pipeline edits with the nodes that changed, deploys, connection changes, backups and settings. Secrets are never recorded.

Download a pipeline as one file, and import one back as a new draft.

Start, follow and cancel runs from AWS Step Functions, EventBridge, Logic Apps or a script.

The pipeline file explained, and every node's PySpark and pandas equivalent.