Workflow - Key Concepts

Overview

In Paygate, workflows allow you to automate tasks so that they run autonomously.

Key concept 1 - Workflows must be invoked by a trigger.

Key concept 2 - Triggers are time based and usually run on a reoccurring schedule.

Key concept 3 - When triggered, a workflow is placed on a queue. Once on the queue, the workflow is guaranteed to run but may not run immediately. The system monitors resources and will automatically run a workflow from the queue when sufficient resources are available.

A basic overview of how Workflows work.

Paygate automation uses the concept of triggers to start the workflow execution process. A workflow attached to a particular trigger is known as a bound workflow.

When a trigger fires, any bound workflows of that trigger will be eligible to be executed. To continue the execution process the trigger will place any bound workflows onto a special queue were they will wait to be picked up by one of many workflow runners.

Key Concept 1

Workflow Runners are self-contained, independent systems that have one purpose - to execute workflows. They do no other work than executing workflows and each runner will only execute one workflow at any one time.

This process of a trigger firing, passing the workflow to the queue and runners taking workflows from the queue is shown in the following diagram:

Key Concept

Workflow Diagrams

In Paygate, workflows are created as diagrams using a tool called the Workflow Designer. The Workflow Designer is available from within the Paygate client using a standard browser.

A workflow Diagram consists of individual blocks called nodes. Each node carries out a specific piece of work such as copying a file or downloading BACS reports.

For example, consider the workflow in the diagram below:

Basic Workflow

The workflow starts at the Start node on the left. It then flows, step-by-step, to the right:

It’s a simple workflow that can save hours of time, each month.

Other concepts

When a Workflow is executed, the workflow engine loads up a workflow diagram and finds the ‘Start Node’, the origin of all workflows. The engine will then follow the arrows in the diagram, one at a time and completing each node before moving onto the next.

Workflow nodes are never run in parallel.