Skip to main content

Workflows

Workflows are the heart of Ecosystem. They define the data processing logic triggered by a trigger.

What is a workflow?​

A workflow is a sequence of operations (nodes) that:

  • Receives data from a trigger
  • Processes this data according to the defined logic
  • Produces a result (HTTP response, email send, etc.)

Structure of a workflow​

Nodes​

Each workflow is made up of nodes connected to each other:

  • Trigger node: entry point (automatic, based on the associated trigger)
  • Processing nodes: conditions, loops, code, etc.
  • Output or action nodes: HTTP request, HTTP response, email send, etc. They are not necessarily terminal: other nodes can follow.

The details of each node type (fields, behavior, use cases) are documented in the dedicated pages — see the node catalog below.

Connections (Edges)​

Nodes are connected by arrows that define the execution flow:

  • Data passes from one node to another
  • The context is shared between all nodes
  • Errors can be handled through specific nodes

Workflow editor​

The workflow editor offers:

  • Graphical interface: drag-and-drop of nodes
  • Visual configuration: configuration of each node through forms
  • Testing: test the workflow on a development environment — start by testing on this type of environment.

The workflow board shows the Last Updated column New : the workflow's business last-edit date.

Node types​

The following pages serve as a reference to configure each node in the editor:

CategoryNodes
HTTP input / outputTrigger node, HTTP(S) Request, HTTP Response
ProcessingCode, AI Agent
CompositionCall a workflow
DatabasesSnowflake, PostgreSQL
Flow controlCondition, Switch, ForEach, While
Cache NewCache nodes
ActionsSend an email, FTP, Send Google Pub/Sub, Send AWS SQS
UtilitiesWait, Note (not executed)
ErrorsOn Error

Workflow lifecycle​

  1. Creation: the workflow is created with the Draft status
  2. Editing: configuration of nodes and connections
  3. Testing: test the deployed workflow on a development environment (test data)
  4. Versioning: work on the Main version; create numbered versions to freeze and release (notably in production)
  5. Deployment: deploy in an environment
  6. Execution: automatic execution as soon as an event is received by the associated trigger

Workflow statuses​

  • Draft: being edited
  • Main version: development line that is always modifiable, ideal for iterating and testing quickly; not deployable in production — see Version management. Deploying other versions requires a trigger on the target environment
  • Deployed: deployed in at least one environment

Next steps​