CODEWITHSUNDEEP
pythonazureazure-functionsfastapiazure-durable-functions
Rohit Kumar S
Rohit Kumar S

Reputation: 11

Can we instantiate more then one top level function app instances

We are trying to create a single Azure Function App in Python that combines a FastAPI framework for building a backend API and Azure Durable Functions for handling long-running processes. In the example code below, a FastAPI application is instantiated on line 7. This application has multiple internal routes and is working as expected, providing a stable and reliable backend API.

We also want to integrate Azure Durable Functions for long-running tasks. As shown on line 9, we're attempting to instantiate a durable function named myApp.

we would also want to know the implication of doing this as durable functions can be blocking actions and fast api would be asynchronous actions mostly.

we tried to split these into two function app and its working fine, we are trying to check if it is possible to run in single function app in Azure.

Upvotes: 0

Views: 149

Answers (1)

Arko
Arko

Reputation: 3781

It is possible to instantiate more than one top-level function app instance in Azure Functions. Each instance can be configured with its own set of functions, pricing plan, deployment method, and runtime version.

Regarding the specific use case of combining FastAPI and Azure Durable Functions in a single function app, it is possible to do so.

You will need a main.py to contain the FastAPI application instance (in my case- app)

from fastapi import FastAPI

  

app = FastAPI()

  

@app.get("/")

async  def read_root():

return {"message": "Hello from FastAPI"}

Create a function app

import json

import azure.functions as func

import azure.durable_functions as DurableFunc

from app.main import app as fastapi_app

  

app = func.AsgiFunctionApp(app=fastapi_app, http_auth_level=func.AuthLevel.ANONYMOUS)

myApp = DurableFunc.DFApp(http_auth_level=func.AuthLevel.ANONYMOUS)

  

@myApp.route("orchestrators/{functionName}")

@myApp.durable_client_input(client_name="client")

async  def http_start(req: func.HttpRequest, client):

function_name = req.route_params.get('functionName')

if req.get_body():

request_body = json.loads(req.get_body().decode())

else:

request_body = None

  

instance_id = await client.start_new(function_name, client_input=request_body)

response = client.create_check_status_response(req, instance_id)

return response

Create a host.json and local.settings.json

host.json

{
    "version": "2.0",
    "extensions": {
        "durableTask": {
            "hubName": "MyTaskHub"
        }
    }
}

local.setting

{
    "IsEncrypted": false,
    "Values": {
        "AzureWebJobsStorage": "UseDevelopmentStorage=true",
        "FUNCTIONS_WORKER_RUNTIME": "python"
    }
}

Create your Durable Functions

import azure.durable_functions as df

def orchestrator_function(context: df.DurableOrchestrationContext):
    result = yield context.call_activity('Hello', "world")
    return result

main = df.Orchestrator.create(orchestrator_function)

Create an Activity Function

import azure.functions as func

def main(name: str) -> str:
    return f"Hello {name}!"

You can verify the same-

uvicorn main:app --reload

enter image description here

You can now deploy the same to your function app

func azure functionapp publish arkoFuncApp

References:

Upvotes: -1

Related Questions