Webhook body parsed as string not object

[esisilvestri]

I have an external system that makes a POST request to a Prefect Webhook containing a JSON object. However when I look at the event in Prefect the body is always just appears as a string.

Based on this information I was under the impression that the body should be parsed as an object:

https://docs.prefect.io/v3/automate/events/webhook-triggers#how-http-request-components-are-handled

Is this the expected behaviour or am I making a mistake somewhere?

The json that is posted has the following structure:

{ "data": "", "deployment_type": "bulk", "file": "filename" }

I have an autoamtion connected ot this webhook and in it I am trying to pull out the individual elements as parameters for the Flow using the Jinja syntax. For example {{ event.payload.body.deployment_type }} for the deployment_type however when I look at the deployment that is run as a result of the automation nothing is returned for this parameter.

One option I can see is to just pass the whole body as one parameter and then use

values = json.loads(data)

but I would prefer to get the body as seperate parameters if possible.

[zzstoatzz]

hi @esisilvestri - this is technically possible today, but the DX is a bit wonky

related to #14875

[esisilvestri]

Hello @zzstoatzz. Thanks for the feedback. I will try this approach. In the meantime I was able to solve it in another way as well. I managed to get the individual elements into the Webhook Template like so:

{
    "event": "webhook.called",
    "resource": {
        "prefect.resource.id": "focus.customers.add",
        "data": "{{ body.data }}",
        "deployment_type": "{{ body.deployment_type}}",
        "file": "{{ body.file}}"
    }
}

Then I was able to access the individual elements in the automation like so {{ event.resource.deployment_type }}

The only issue I still had after that was that data can also be a JSON object and with the above approach that comes in as a string that cannot be lloaded with json.loads(). I was however able to use the "ast" libarary to load the resulting string into a Python dictionary and access the values in my Flow: new_values = ast.literal_eval(data)