Pipelines - Overview

Pipelines stream data continuously between a source and destination, transforming with Vyne along the way


Overview

Where queries in Vyne are a great for on-demand data transformation, Pipelines provide continuously streaming data between source and destination, using Vyne to transform data along the way.

./pipelines-overview.png

In beta

Pipelines are a new addition to the Vyne stack. We're making them available as an early preview release, but you may hit a few rough edges.

Reach out to us on Slack to discuss any problems you encounter, or to give feedback

Pipelines work by reading content from a source, transforming using Vyne, then writing to a Sink.

Pipelines can be defined either in a config file, or using the UI.

Starting Pipeline Engine

The pipeline engine ships as a separate docker image.

A typical docker-compose config looks as follows:

version: "3.3"
services:
  ## Other services omitted
  pipeline-engine:
    image: vyneco/pipeline:${VYNE_VERSION}
    depends_on:
      - schema-server
      - eureka
      - vyne
    environment:
      OPTIONS: >-
        --eureka.uri=http://eureka
        --vyne.schema.consumptionMethod=RSOCKET
        --vyne.pipelines.pipeline-path=./pipeline-config
        --vyne.connections.config-file=./config/connections.conf
        --server.port=80

Application config

The pipeline server can be configured with the following properties.

These are generally passed into the environment.OPTIONS block within a docker services item in docker compose.

Config ParameterDescription
vyne.pipelines.pipeline-pathSpecifies a directory containing multiple pipeline spec .pipeline.json files. Defaults to ./pipelines
vyne.connections.config-filePath to the config file of connections. Should be the same file as passed to Vyne. Defaults to config/connections.conf
vyne.schema.consumptionMethodDefines how to connect to the schema server and consume schema updates. RSOCKET is a reasonable default

Creating a pipeline

Pipelines are created and stored as a series of *.pipeline.json files, written to the configured directory.

Pipelines are defined as an Input (transport + spec) and one or more Outputs (transport + spec), with Vyne handling the transformation and enrichment.

Pipelines listen on the input, and as messages arrive they are passed to Vyne to transform to the output models.

The data is transformed, and Vyne's query engine is used to enrich and populate any missing data.

Although all pipelines are persisted as json files, they can be created either through the UI, or authored as json directly.

Creating a pipeline using the UI

The Vyne WebUI provides a wizard for building pipelines.

Click on the Pipeline Manager in the left-hand navigation bar, and complete the form to create a pipeline.

Creating a pipeline through configuration

Pipelines are authored and stored in json files, with a file per pipeline.

Files should adhere to the pipeline spec, which has its own dedicated reference documentation.