Skip to main content

Message Service

Purpose​

The Message Service is the producer side of layline.io's internal publish/subscribe messaging system. It publishes messages to topics defined in a Message Source. Messages can be published from JavaScript or Python processors, making it useful for:

  • Triggering downstream Workflows based on events
  • Broadcasting data to multiple consumers simultaneously
  • Decoupling processing stages across Workflows

Architecture​

The Message Service is the counterpart to the Message Source (consumer). Together they form a publish/subscribe system:

A single Message Service can reference multiple Message Sources. For each Message Source, you define functions that publish to specific topics.

This Asset can be used by:​

Asset typeLink
ProcessorsJavaScript Processor
Python Processor
AssetDescription
Message SourceDefines the topics that this service publishes to

Configuration​

Name & Description​

  • Service Name : Name of the Asset. Spaces are not allowed in the name.

  • Service Description : Enter a description.

The Asset Usage box shows how many times this Asset is used and which parts are referencing it. Click to expand and then click to follow, if any.

Required Roles​

In case you are deploying to a Cluster which is running (a) Reactive Engine Nodes which have (b) specific Roles configured, then you can restrict use of this Asset to those Nodes with matching roles. If you want this restriction, then enter the names of the Required Roles here. Otherwise, leave empty to match all Nodes (no restriction).

Sources​

Under Sources, add references to the Message Sources whose topics this service publishes to. Each entry links a Message Source asset to this service.

ColumnDescription
SourceSelect a Message Source asset from the project

Click Add source to add a new Message Source reference.

Functions​

Functions (Message Service)

Functions define the publish operations available to this service. Each function maps to a topic in the referenced Message Source(s).

Click Add Function to create a new function.

Function Name & Description​

  • Function name : Name of the function. Must be unique within the service. Must not contain spaces.

  • Description : Optional description of the function.

Request Type​

  • Request type : The data dictionary type of the message payload that will be published. This defines the structure of the data sent when calling this function.

Response Type​

  • Response type : The data dictionary type of the response (if any). Optional. If not set, the function returns no response.

Auto-Generated Function Names​

Message Service functions are auto-generated and invoked from processors using the pattern:

services.<ServiceName>.<FunctionName>

For example, a service named OrderMessageService with a function named PublishOrder would be called as:

services.OrderMessageService.PublishOrder({ ... })

Function Parameters​

When calling a Message Service function from a processor, the following parameters are available:

ParameterDescription
TopicThe name of the topic to publish to (as defined in the Message Source)
PartitionKeyOptional. Used to determine which partition the message goes to (for ordering guarantees within a key)
RequestThe message payload, matching the Request type defined above
SourceOptional. The name of the Message Source to publish to. Required if the service references more than one Message Source

Data Dictionary​

The Data Dictionary allows you to define complex data structures which can be mapped onto external data types and vice versa. This is necessary whenever an asset needs to exchange structured data with an external system — for example, when reading from or writing to a database, an HTTP API, a message queue, or any other format that carries typed fields.

Rather than hard-coding external field names and types into your Workflow, you define your own internal data types here. These internal types are then mapped to the external system's fields at the asset level. This means your Workflow scripts work with consistent, self-documenting data structures regardless of which external system the data came from.

When you need it​

Whenever you configure an asset that exchanges structured data — a JDBC Service, a DynamoDB Service, an HTTP endpoint, an MQ message, a database Resource — you use the Data Dictionary to declare the types that represent:

  • Request parameters — the data your Workflow sends to the external system
  • Result data — the data the external system returns to your Workflow
  • Intermediate structures — types that hold data during a transformation

Entity Types​

The Data Dictionary is organized as a tree of typed entities. The available entity types are:

EntityDescription
NamespaceGroups related types. Optional. If you reuse a namespace name that already exists in the Project, the two namespaces merge.
SequenceAn ordered list of typed members. Members are accessed by name, e.g. MyNamespace.Customer.Name.
EnumerationA fixed set of named integer constants.
ChoiceA type that holds exactly one of several possible member types.
ArrayA sequence of elements of a single contained type.

Defining Types — Step by Step​

The following walkthrough shows how to build a data structure using the Data Dictionary editor. The example assumes a SQL customer table with columns id, name, and address — but the same pattern applies whenever you need to declare types for any asset.

1. Declare a new type​

Click Declare Root Type in the toolbar to add a top-level entity.

Declare root type

2. Declare a namespace (optional)​

Namespaces organize related types. To add one, right-click an existing node and select Add Sibling, then set the element type to Namespace.

Declare namespace

  • Name — The name of the namespace. If a namespace with this name already exists elsewhere in the Project, their contents merge automatically. Otherwise the name must be unique and may not contain spaces.

  • Type — Pick the entity type. For a namespace, select Namespace.

  • Description — Optional free-text description.

3. Declare a Sequence under the namespace​

Right-click the namespace and choose Add Child to add a child element.

Add child to namespace

Click the arrow next to the namespace name and select Add child. Then fill in the element details:

Declare sequence

  • Name — The name of the element, e.g. Customer.

  • Type — Select Sequence as the element type. You will add individual fields (members) in the next step.

  • Extendable Sequence — When checked, layline.io can dynamically extend the sequence's member list if incoming data contains fields that are not explicitly defined. Leave unchecked if all fields are known in advance.

4. Add members to the Sequence​

With the Sequence selected, click Add Child to add individual fields:

Add sequence members

Each member maps to a column in the external data source. You can reference any member by its full path — for example, MyNamespace.Customer.Name — from your Workflow scripts.

Common Entity Fields​

These fields are available on all entity types:

FieldDescription
NameUnique identifier within the namespace. Reusing a namespace name from another part of the Project merges the two.
TypeThe entity kind: Namespace, Sequence, Enumeration, Choice, or Array
DescriptionOptional free-text description
Extendable Sequence(Sequence only) Allows the member list to be extended dynamically at runtime
Members(Sequence) Ordered list of typed fields — click Add Child to add each one
Elements(Enumeration) Named integer constants making up the enumeration

Advanced Features​

Inheritance and Override
Entities inherited from a parent format or resource appear in the tree in a distinct inherited style. These are read-only unless overridden. Click Reset to Parent on an overridden entity to restore the inherited definition.

Copy and Paste
Use the toolbar buttons to copy a complete entity subtree and paste it elsewhere in the tree. All members and nested entities travel with it.

Filter and Sort
Use the Filter field to search entities by name. The sort buttons order nodes ascending or descending alphabetically.

See Also​

Using the Message Service from a Script Processor​

Publishing a Message​

/**
 * Publish an order confirmation message
 * @param orderData Order confirmation data
 */
function publishOrderConfirmation(orderData) {
    services.OrderMessageService.PublishOrderConfirmation({
        Topic: "order-confirmations",
        PartitionKey: orderData.orderId,
        Request: orderData
    });
}

Publishing with Automatic Source Selection​

If the service references only one Message Source, the Source parameter can be omitted:

/**
 * Publish a notification event
 * @param eventData The event to publish
 */
function publishNotification(eventData) {
    services.NotificationService.SendNotification({
        Topic: "notifications",
        Request: eventData
    });
}

Publishing with Explicit Source​

If the service references multiple Message Sources, specify which one to use:

/**
 * Publish to a specific Message Source
 * @param data The data to publish
 */
function publishToSpecificSource(data) {
    services.MultiSourceService.PublishData({
        Source: "CustomerDataSource",
        Topic: "customer-updates",
        PartitionKey: data.customerId,
        Request: data
    });
}

Full Example: Order Processing Pipeline​

A complete example showing how Message Source and Message Service work together:

  1. Define a Message Source named OrderSource with a topic order-confirmations
  2. Define a Message Service named OrderMessageService that references OrderSource
  3. Create a function PublishOrderConfirmation with a Request type of OrderConfirmation
  4. In a script processor, call the service to publish when an order is confirmed:
/**
 * Called when an order has been confirmed
 * @param orderId The ID of the confirmed order
 * @param orderDetails The order details
 */
function onOrderConfirmed(orderId, orderDetails) {
    let confirmationData = dataDictionary.createMessage(
        dataDictionary.type.OrderConfirmation
    );
    confirmationData.data.orderId = orderId;
    confirmationData.data.customerId = orderDetails.customerId;
    confirmationData.data.timestamp = DateTime.now().toString();

    services.OrderMessageService.PublishOrderConfirmation({
        Topic: "order-confirmations",
        PartitionKey: orderId,
        Request: confirmationData
    });
}

A downstream Workflow that references OrderSource will receive this message via its Input Processor and can process the confirmation further.

For more on processors, see JavaScript Processor and Python Processor.

Service Testing​

layline.io provides a test facility for testing your Services before you deploy them. In this way, you save time and effort by testing your Services without having to deploy and activate a whole Project with Workflows.

Once you have configured your Service(s), you can test them: Within your Asset Configuration tab (1), switch to the Test tab (2) to test your Service.

Test Facility Toolbar​

The toolbar provides the following options:

The Testing tab provides two major views:

  1. Testcase configuration: This is where you define the testcases to be executed.
  2. Testcase execution: This is where you can execute the testcases and see the results.

You switch between these two views by clicking on the leftmost icon in the toolbar (1).

Let's start with the Testcase configuration view.

Testcase Configuration​

The concept of the Testing is to define a set of Testcases which can be executed in a batch or individually. For this purpose, you can define multiple Testcases and configure them individually. I.e. each Testcase groups a number of indidivual tests which can be executed individually or in a batch.

Adding a Testcase​

Click Add Testcase in the toolbar to add a new testcase:

A new Testcase is added. It is automatically named New<Service Asset Name>Test (3) and added to the list of Testcases (2).

  • Service test name (3): You can change the name of the Testcase here.
  • Service test description (4): You can add a description to the Testcase here.

Test Case Setup​

Basics​

In this section you define the individual tests to be executed for this Testcase.

To start, click # END in the toolbar:

A new test is added to the list of tests (1), and the test is opened for configuration (2).

Next we fill in the details:

  • Test name (3): You can change the name of the Test here.

  • Test description (4): You can add a description to the Test here.

  • Service function to test (5): Select the Service function to test here.

    This list contains all Service functions which are defined in the Service Asset. Pick the one you want to test.

    Once a Service function is selected, the system will automatically create a skeleton to fill in the respective parameters for the selected Service function.

Service Function Input Parameters​

  • Service Function Input Parameters (6): Fill in the respective parameters for the selected Service function.

    In our example we have a function GetAlertsForSite which takes two parameters baseurl and riskId. If we click on Add member in the skeleton table the system will allow you to select the respective parameter from the list of available parameters:

    Once you have selected the parameter, the system will automatically add the respective parameter name. You then add the respective value for the parameter:

Service Function Evaluation Parameters​

To automatically evaluate the result, you can add a script which analyzes the results.

Testcase Execution​

Once you have configured your Testcases, you can execute them.

There are two ways on how to trigger execution:

  • Option 1: Select Run selected test in the toolbar (1) to execute the currently selected Testcase.

    Executing a test this way will switch the tab to the Testcase execution view, execute the test and show the results.

  • Option 2: Switch to the Testcase execution view by clicking on the leftmost icon in the toolbar (1) select the test to execute, and then hit the play button next to the test.

Each option will take us to the Testcase execution view:

In this view you can find the Testcase (1) and the Tests (2) we have created. If we had created additional tests for this Testcase, they would be listed here as well.

Question marks indicate that the test has not yet been executed.

We can now either execute all tests, or run them individually:

  • Run all Tests (1): Click this button to execute all tests.

  • Run Testcase (2): Click this button to a Testcase with all its underlying individual tests.

  • Run individual Test (3): Click this button next to a test to execute this individual test.

Once a test has been executed, the question mark will be replaced by a green check mark or a red cross depending on whether the test was successful or not.

The right hand-panel will show the results of the test execution respectively:

In case of errors, the system will show the error message for further investigation.

Can't find what you are looking for?

Please note, that the creation of the online documentation is Work-In-Progress. It is constantly being updated. should you have questions or suggestions, please don't hesitate to contact us at support@layline.io .