Webhook Sink¶

The Webhook Sink block enables sending a data from Workflow into external APIs by sending HTTP requests containing workflow results.

How This Block Works¶

It supports multiple HTTP methods (GET, POST, PUT) and can be configured to send:

• JSON payloads

• query parameters

• multipart-encoded files

This block is designed to provide flexibility for integrating workflows with remote systems for data exchange, notifications, or other integrations.

Setting Query Parameters¶

You can easily set query parameters for your request:

query_parameters = {
    "api_key": "$inputs.api_key",
}

will send request into the following URL: https://your-host/some/resource?api_key=<API_KEY_VALUE>

Setting headers¶

Setting headers is as easy as setting query parameters:

headers = {
    "api_key": "$inputs.api_key",
}

Sending JSON payloads¶

You can set the body of your message to be JSON document that you construct specifying json_payload and json_payload_operations properties. json_payload works similarly to query_parameters and headers, but you can optionally apply UQL operations on top of JSON body elements.

Let's assume that you want to send number of bounding boxes predicted by object detection model in body field named detections_number, then you need to specify configuration similar to the following:

json_payload = {
    "detections_number": "$steps.model.predictions",
}
json_payload_operations = {
    "detections_number": [{"type": "SequenceLength"}]
}

Multipart-Encoded Files in POST requests¶

Your endpoint may also accept multipart requests. You can form them in a way which is similar to JSON payloads - setting multi_part_encoded_files and multi_part_encoded_files_operations.

Let's assume you want to send the image in the request, then your configuration may be the following:

multi_part_encoded_files = {
    "image": "$inputs.image",
}
multi_part_encoded_files_operations = {
    "image": [{"type": "ConvertImageToJPEG"}]
}

Cooldown¶

The block accepts cooldown_seconds (which defaults to 5 seconds) to prevent unintended bursts of notifications. Please adjust it according to your needs, setting 0 indicate no cooldown.

During cooldown period, consecutive runs of the step will cause throttling_status output to be set True and no notification will be sent.

Async execution¶

Configure the fire_and_forget property. Set it to True if you want the request to be sent in the background, allowing the Workflow to proceed without waiting on data to be sent. In this case you will not be able to rely on error_status output which will always be set to False, so we recommend setting the fire_and_forget=False for debugging purposes.

Disabling notifications based on runtime parameter¶

Sometimes it would be convenient to manually disable the Webhook sink block. This is possible setting disable_sink flag to hold reference to Workflow input. with such setup, caller would be able to disable the sink when needed sending agreed input parameter.

Common Use Cases¶

• Use this block to [purpose based on block type]

Connecting to Other Blocks¶

The outputs from this block can be connected to other blocks in your workflow.

Type identifier¶

Use the following identifier in step "type" field: roboflow_core/webhook_sink@v1to add the block as as step in your workflow.

Properties¶

The Refs column marks possibility to parametrise the property with dynamic values available in workflow runtime. See Bindings for more info.

Runtime compatibility¶

requires_internet — air-gapped / offline deployments

This block depends on a service that is not reachable from fully offline / air-gapped deployments.

soft — runtime hosted_serverless, dedicated_deployment; execution remote

Cooldown / rate-limit timer is stored in process memory. With remote step execution on stateless or multi-replica HTTP runtimes each request gets a fresh worker, so cooldown does not throttle. Cooldown only behaves as documented with local step execution inside an InferencePipeline.

Available Connections¶

Input and Output Bindings¶

The available connections depend on its binding kinds. Check what binding kinds Webhook Sink in version v1 has.

{
    "name": "<your_step_name_here>",
    "type": "roboflow_core/webhook_sink@v1",
    "url": "<block_does_not_provide_example>",
    "method": "<block_does_not_provide_example>",
    "query_parameters": {
        "api_key": "$inputs.api_key"
    },
    "headers": {
        "api_key": "$inputs.api_key"
    },
    "json_payload": {
        "field": "$steps.model.predictions"
    },
    "json_payload_operations": {
        "predictions": [
            {
                "property_name": "class_name",
                "type": "DetectionsPropertyExtract"
            }
        ]
    },
    "multi_part_encoded_files": {
        "image": "$steps.visualization.image"
    },
    "multi_part_encoded_files_operations": {
        "predictions": [
            {
                "property_name": "class_name",
                "type": "DetectionsPropertyExtract"
            }
        ]
    },
    "form_data": {
        "field": "$inputs.field_value"
    },
    "form_data_operations": {
        "predictions": [
            {
                "property_name": "class_name",
                "type": "DetectionsPropertyExtract"
            }
        ]
    },
    "request_timeout": "$inputs.request_timeout",
    "fire_and_forget": "$inputs.fire_and_forget",
    "disable_sink": false,
    "cooldown_seconds": "$inputs.cooldown_seconds"
}