[Yandex Cloud documentation](../../../index.md) > [Yandex Cloud Functions](../../index.md) > [Developing in Python](index.md) > Request handler # Request handler for a function in Python A _request handler_ is a method used to process each Python function call. When creating a function [version](../../concepts/function.md#version), you must set up an _entry point_ for it, i.e., a path to the request handler in ``.`` format, where: * ``: Name of the file with the function code (without `.py`), e.g., `index`. The code file must reside in the root directory. The file name must not contain any dots. * ``: Name of the callable object in ``, e.g., `handler`. When initializing the function, the runtime imports `` and finds the callable object named `` which runs on every function invocation. Example of an entry point for a Python function: `index.handler`. {% note info %} At any given time, a single function instance processes only one request. This allows you to use global variables without the need to ensure data integrity control. {% endnote %} When calling the handler, the runtime provides the following arguments: 1. Request body (the `event` parameter): * If the request body is a [JSON document](../../concepts/function-invoke.md#request), it will be converted into `dict` using the `json.loads` method. * If the function was invoked with the `?integration=raw` query parameter, the HTTP request body is provided to the function as is, i.e., unprocessed. 1. [Invocation context](context.md) (the `context` parameter). The context provides all required information about the function version. The structure of this object is described in [Python function invocation context](context.md). ## Handler types {#type} A function supports both synchronous and asynchronous handlers. ## Synchronous handler {#sync} To return the execution result, use the `return` statement or raise an exception using the `raise` statement. A synchronous function must return a result or raise an exception. ## Asynchronous handler {#async} A handler can be an `async def` asynchronous function. In this case, you can use the following statements: * `return`: Returns the function response. * `raise`: Reports an error to the runtime. * `await`: Tracks the execution of asynchronous function invocations. {% note info %} The `asyncio` library is the only supported runtime for asynchronous functions. In Python 3.14 and higher, `asyncio.get_event_loop()` no longer automatically creates an [event loop](https://docs.python.org/3/library/asyncio-eventloop.html), but results in `RuntimeError`. Use `asyncio.get_running_loop()` to manage asynchronous handlers. See [this example](#async-python314-example). {% endnote %} For more information about the development process using `async/await`, see [this article](https://docs.python.org/3.14/library/asyncio.html). ## Examples {#examples} ### HTTP request structure output {#http-req} The following function outputs the request structure and invocation context to both the execution log and function response: ```python import json def handler(event, context): return { 'statusCode': 200, 'body': json.dumps( { 'event': event, 'context': context, }, default=vars, ), } ``` Function invocation example: ```bash curl \ --data '{"hello": "world"}' \ --header 'Content-Type: application/json' \ https://functions.yandexcloud.net/?param=one ``` Result: ```bash { "context": { "aws_request_id": "6e8356f9-489b-4c7b-8ba6-c8cd********", "deadline_ms": 1657713543198, "function_name": "d4eo2faf62**********", "function_version": "d4e3vrugh3**********", "invoked_function_arn": "d4eo2faf62**********", "log_group_name": "ckgjmanjlh**********", "log_stream_name": "d4e3vrugh3**********", "memory_limit_in_mb": 128, "request_id": "6e8356f9-489b-4c7b-8ba6-c8cd********", "token": { "access_token": "", "expires_in": 42299, "token_type": "Bearer" } }, "event": { "body": "{\"hello\": \"world\"}", "headers": { "Accept": "*/*", "Content-Length": "18", "Content-Type": "application/json", "Host": "functions.yandexcloud.net", "User-Agent": "curl/7.64.1", "X-Forwarded-For": "109.252.148.209", "X-Real-Remote-Address": "[109.252.148.209]:2816", "X-Request-Id": "6e8356f9-489b-4c7b-8ba6-c8cd********", "X-Trace-Id": "e9fe9b05-c1aa-4fb8-94d8-a514********" }, "httpMethod": "POST", "isBase64Encoded": false, "multiValueHeaders": { "Accept": [ "*/*" ], "Content-Length": [ "18" ], "Content-Type": [ "application/json" ], "Host": [ "functions.yandexcloud.net" ], "User-Agent": [ "curl/7.64.1" ], "X-Forwarded-For": [ "109.252.148.209" ], "X-Real-Remote-Address": [ "[109.252.148.209]:2816" ], "X-Request-Id": [ "6e8356f9-489b-4c7b-8ba6-c8cd********" ], "X-Trace-Id": [ "e9fe9b05-c1aa-4fb8-94d8-a514********" ] }, "multiValueParams": {}, "multiValueQueryStringParameters": { "param": [ "one" ] }, "params": {}, "pathParams": {}, "queryStringParameters": { "param": "one" }, "requestContext": { "httpMethod": "POST", "identity": { "sourceIp": "109.252.148.209", "userAgent": "curl/7.64.1" }, "requestId": "6e8356f9-489b-4c7b-8ba6-c8cd********", "requestTime": "13/Jul/2022:11:58:59 +0000", "requestTimeEpoch": 1657713539 }, "url": "" } } ``` ### Using an asynchronous handler in Python 3.14 {#async-python314-example} The following function demonstrates how to use an asynchronous handler in Python 3.14 and higher, where explicit event loop management is required: ```python import asyncio import json def handler(event, context): loop = asyncio.new_event_loop() asyncio.set_event_loop(loop) async def async_handler(): current_loop = asyncio.get_running_loop() current_time = current_loop.time() return { 'statusCode': 200, 'body': json.dumps({ 'message': 'Async handler with explicit loop creation', 'loop_time': current_time, 'python_version': '3.14', }) } return loop.run_until_complete(async_handler()) ``` Function invocation example: ```bash curl \ --data '{"test": "asyncio"}' \ --header 'Content-Type: application/json' \ https://functions.yandexcloud.net/ ``` Result: ```json {"message": "Async handler with explicit loop creation", "loop_time": 3.522407882, "python_version": "3.14"} ```