Skip to content
On this page

Flows

Flows enable custom, event-driven data processing and task automation within Directus. Each flow is composed of one trigger, followed by a series of operations.

Before You Begin

Please be sure to see the Quickstart Guide to get a basic overview of the platform.

Learn More

There is also dedicated API documentation on Flows and Operations.

What's a Flow?

What's a Flow?

Each flow is made up of three elements: A trigger, operations, and a data chain.

Triggers

Each flow begins with a trigger, which defines the action or event that starts the Flow. This action or event could be some type of transaction within the app, an incoming webhook, a cron job, etc.

Operations

An operation is an action or process performed within the flow. These enable you to manage data: send off emails, push in-app notifications, send webhooks, and beyond.

To put it in conceptual terms, operations do three things:

  • Get data from Directus or another outside service.
  • Process data a.k.a. transform it, validate it, or whatever.
  • Send data to Directus or another outside service.

Developers

You can even develop your own custom operations to fit any use-case.

Data Chains

In order for a flow's operations to track and access the same data, each flow creates its own data chain. Every operation has access to this data chain and each operation appends some value onto this object after it runs. This means you can dynamically access data from a previous operation in the current operation with data chain variables.

Control Flow

Not every operation that executes in a flow does so successfully. In some cases, your operations are going to fail. Perhaps an operation tried to access data that doesn't exist, or a webhook operation fails for some reason, or perhaps you set a condition operation, which fails by design when its condition is not met.

These kinds of failed operations do not immediately stop your flow. Instead, flows let you implement control flow, by providing success paths and failure paths within a flow:

  • Success — If operation1 executes successfully, then operation2 executes.
  • Failure — Else if operation1 fails on execution, then operation3 executes.

And there we have it! These are the conceptual cornerstones of any flow. Now you'll need to know how to actually create a flow, which we discuss in the next section.

Configure A Flow

Create a Flow

  1. Navigate to Settings > Flows and click add in the page header. A drawer will open.
  2. Under Flow Setup, fill in a Name for the flow and the following optional details:
    • Status — Sets the flow to active or inactive.
    • Icon — Adds an icon to help quickly identify the flow.
    • Description — Sets a brief verbal description of the flow.
    • Color — Sets a color to help identify the flow.
    • Activity and Logs Tracking — Lets you Track Activity and Logs, Activity, or Neither.

TIP

To learn more, see the section below on Logs as well as the Activity Log documentation.

Configure a Trigger

  1. Click arrow_forward to navigate to Trigger Setup. Select a trigger type and configure as desired.
  2. Click done in the menu header to confirm.

You'll now see your trigger in an empty grid area. Its time to start adding operations.

Configure an Operation

  1. On the trigger panel, click add and the Create Operation side drawer will open.
  2. Choose a Name, an operation type, and configure as desired.
    Directus will convert the name into a unique operation key, used on the data chain.
    If you don't choose a name, the system will auto-generate a name and key for you.
  3. Next, click done in the page header to confirm and return to the flow grid area.
  4. From here, you can make the following optional configurations:
    • Reposition — You can drag and drop panels to reposition as desired.
    • Unlink/Relink — Click and drag adjust or arrow_forward to unlink/relink flows.
    • Duplicate an Operation — To duplicate an operation, click more_vert to open its context menu. Click control_point_duplicate and a duplicate of the operation (and its configuration details) will be created.
    • Copy an Operation — To copy and paste an operation into another flow, click more_vert to open its context menu. Click input and a popup menu will open. Choose the desired flow from the dropdown and click Copy.
    • data_object Toggle Raw Editor — Click data_object on an operations' form input fields to toggle the input type between standard and raw value. This is allows you to add a raw value or Data Chain Variables within any type of configuration option, even dropdown menus, checkboxes, and radio buttons.
    • Delete an Operation — To delete an operation, click more_vert then delete. A popup menu will appear. Click Delete to confirm.
  5. On the newly created operation panel:
    • Click add to add an operation to the success path.
    • Click remove to add an operation to the failure path.
  6. Repeat steps 5-10 to build out your flow as desired.
  7. Click done to confirm and create your flow.
  8. Click arrow_back to return to the flows list.
  9. Once created, you may need to re-edit your flow, toggle it to inactive, or delete it.

Edit a Flow

  1. Navigate to the desired flow.
  2. Click edit in the flow page header and make reconfigurations as desired.
  3. Click done to confirm.

Toggle a Flow to Inactive

  1. Navigate to Settings > Flows and click more_vert on the desired flow.
  2. Click check Set Flow to Active or block Set Flow to Inactive.

Delete a Flow

  1. Click more_vert on the desired flow to open its context menu.
  2. Click delete and a popup menu will appear. Click Delete to confirm.

Now that we know how to create and configure a flow, it's time to get a firmer understanding of the data chain.

The Data Chain

Remember, each flow creates its own JSON object to store any data generated.

When the flow begins, three keys are appended to the data chain: $trigger, $accountability, and $last. Then, as each operation runs, it has access to this data chain. Once an operation finishes, its data is appended under its <operationKey>. When the operation doesn't generate data, null is appended under its key.

The following is a highly generic example of a data chain.

json
{
	"$trigger": {
		// Contains data generated by the flow's trigger.
		// This could include headers, access tokens, payloads, etc.
		// Every data chain has a $trigger key.
	},
	"$accountability": {
		// Provides details on who/what started the flow.
		// This could include user's id, role, ip address, etc...
		// Every data chain has an $accountability key.
	},
	"$last": {
		// The value appended under $last changes after each operation.
		// It stores data of the last operation that executed in the flow.
		// That way, you don't have to remember the previous operation's unique keyname.
		// It's a handy little convenience tool!
		// Every data chain has a $last key.
	},
	"operationKey1": "A value", // The data (if any) generated by the first operation.
	"operationKey2": {
		"nestedKey": ["nested val", "nested val 2"] // It will be common to have nested JSON data.
	},
	"operationKey3": null // A null value is appended if no data generated.
}

As you can see, the example above doesn't have any substantial data inside each key. In reality, there's going to be a lot of data and it will always be slightly different, based on your flow's unique configuration. During configuration and debugging, you'll need to use a tool like The Log to view your data chain and make sure each operation is accessing and generating data as you intended.

TIP

In our examples, we are using generic placeholders for operation keys, like <operationKey>, which might look funny to low-code users. In practice, operation keys will actually have unique and descriptive names, like send_email_7538.

TIP

Remember, $trigger, $accountability, and $last begin with $, but not operationKeys.

Data Chain Variables

While configuring your operations, you can use keys from the data chain as variables to access data. Simply wrap the variable with quotes and double mustaches. For example:

json
"{{ $accountability }}"

will get the data nested under the $accountability key, producing something like this:

json
{
	"user": "4b11492d-631d-4b8a-bca7-2beasdfadf58b",
	"role": "12c79228-5361-4905-929b-d69d3w46trs6b",
	"admin": true,
	"app": true,
	"ip": "127.0.0.1",
	"userAgent": "Amazon CloudFront"
}

You can mix your own hard-coded JSON alongside variables.
You can also use dot-notation and array indexing to retrieve sub-nested values.

json
{
	"key0": "a hard-coded value",
	"key1": "{{ $trigger.payload }}",
	"key2": "{{ operationKey.payload.friend_list[0] }}"
}

You cannot pass any type of computation using double-moustache syntax.

js
{
	"key": "{{ 2 + 2 }}",
	"key2": "{{ $trigger.payload.toLowerCase() }}"
}

TIP

To perform computations on flow data, use the script operation or a webhook.

Certain operations use dropdowns, toggles, checkboxes, and other input options. However, you can bypass this entirely to input raw values directly with Toggle to Raw Editor. You can use double-moustache syntax to access data dynamically in these input options as well.

Logs

Accessible from the sidebar, logs store information for each flow execution. Each log will display information from triggers as well as each operation in the flow. To access a flow's logs, follow these steps.

  1. Navigate to Settings > Flows and click the desired flow.
  2. Click fact_check Logs in the sidebar. A side drawer will open, displaying the flow's logs.
  3. Click a log and another side drawer will open, allowing you to peer through its data.
  4. When finished, click close to close the drawer.

Logs are not a 1:1 mapping to the data chain. Each trigger and operation gets its own dropdown, which stores its relevant data. Here's what you'll get from each of these:

Trigger

  • Options — The values you input when you configured the trigger.
    (These configuration options are not stored on the data chain).
  • Payload — Displays the data appended under $trigger.
  • Accountability — Displays data appended under $accountability.

Note that $accountability is not nested under the $trigger key. However, it is listed under the Trigger in the Log because $accountability is generated by the trigger.

<OperationKey>

  • Options — The values you input when you configured the operation.
    (These configuration options are not stored on the data chain).
  • Payload — Displays the data appended under this <operationKey>.

Remember, the Log to Console operation is a key debugging tool. It does not append data to the data chain. You will view your log message under Options. Therefore, anything you log will always be displayed as nested under a message key. For example, if you decide to log "The last operation was a success", it will be displayed as:

{
	"message": "The last operation was a success"
}

Logs are stored in the database

Keep in mind that if you've configured a flow to track logs, all this information is stored in the database. You may need to periodically delete this data.

Where is $last?

You may notice $last is not in the Logs. Remember, $last constantly updates to store the data of the most recently executed operation. The log shows the results of the entire flow. Therefore $last would simply be the very last operation in the flow.

More on Debugging

You may find a tool like Postman quite helpful for viewing data and debugging flows.