Basics

Automation is about predefined actions that transform data from one point to another.

Core concepts

Trigger

Every workflow starts with a trigger. They start the automation. Trigger only have an output branch. They launch and activate the workflow. There are: -manual trigger -Scheduled Trigger - Every minute/hour etc. -Applications - Webhook, Property update, Form submission('n8n Form Trigger')

We can have multiple triggers.

Filter

Filter allow or block certain types of data from following a path based on certain conditions. 'Filter Node'

Actions

Actions allow to interact with applications

Workflow

Trigger --> Sorting/Filtering/Fromatting/Transforming/Segmenting --> Action

First step is to map out the process by creating a flowchart: - List every single step as a block; Get website link -> create image -> save image

API's and Webhooks

API = Application programming interfaces

An API exposes a service and developers write programs to consume it.

In the documentation of an API we see how it works.

We send a request through the interface to the Application -> the application uses the interface to send a response

    --> Request

client Server/Interface/Application <-- Response

For that we for example use an HTTP Request Node

Credentials

Let the application know that we are allowed to make a given request. Most API's require authentification through credentials.

Main ways to authenticate: - Query Parameter: ?api_key=... - Header: Authorization:Bearer...

Webhooks

Webhooks or Reverse API's

Webhooks indicate that something that you are waiting for has happened.

For example: Stripe Everytime a customer made a payment to your stripe account

There are two options to retrieve that information: - Polling - every x minutes check if new payment - Webhook sends when new payment

Nodes

Nodes are the building blocks of n8n

There are 3 main categories of nodes - Entry piont like Triggers - Functions to transform, filter or format data - Exit point - Apps/Application

Types are: - Trigger - Action in App - Data transformation - Flow - Files - Advanced.

In the Node View the left side shows the input data and the right side shows the output data.

Each node executes once per item.

In Output Fieldnamecan we rename the output data key name.

We also can set Execute Once in the node setting - the whole node just executes for first item in input.

Some nodes ask you for field names - in this case you don't use the expression. You use the name

Data and Expressions

We can drag the key from the input field into the nodes fields. That creates an expression "return for each item the associated value for that key" Everything between the {{}} is an expression.

Expression can combine plain text, item variables and javascript.

{{$JSON.name}} says {{$JSON.greeting}}

To create a key:value pair you assign a key name under Namefield and then drag the expressions into the field.

We can add multiple fields under Add Field

Branches

We create branches when we want to set different paths/sets of actions.

Nodes with multiple outputs have two output options to start branches on the right.

Useful Nodes

Execution Logs

We can access the history of all our workflows executions

Error handling

When the workflow fails there will be a Error Workflow executed. It can be configured per workflow in the settings tab. It informs you that your workflow failed.

Anothe way to trigger the error workflow is with the Stop and Error Node.

The error workflow is created with the Error Trigger Node. It contains information about the error: - workflow name & id - execution ID - Name of node with the error - Link to execution history.

It is recommended to notify via Slack, Gmail, Whatsapps/Telegram to send message to these tools when you have an error.

Makes sense for example to do error handling on inputs that come in to the workflow. Like checking if the email is actually valid that came from a webhook.

Debugging

When a workflow fails it can be for multiple reasons: - Node not configured correctly - Service not available - Wrong data format - Missing data

Sometimes a workflow can not fail but also not be successful.

Debug in editor

Easiest way to debug is the Debug in editor feature. It lets you pin data from an execution history into the editor. So it lets you edit the workflow using the data that caused an error.

Once the workflow is fixed we can use the retry feature.

Edit output

You can manually edit the output of specific node. This can be used for testing you workflow.

Workflow Version Hisotry

Access all the historical versions of a given workflow.

Comments