async await inside aws lambdas using python

When working with AWS Lambda and asyncio in a Python environment we can sometimes get some odd behavior with the lambda itself running the asynchronous code but not returning the results.

This is expected because we're not necessarily telling the Lambda to await the asynchronous invocation's results.

Pitfalls with asyncio in AWS Lambdas

A few gotcha's exist when working with asyncio, below are the pitfalls:

Do not install asyncio using pip - The asyncio library is included inside

the Python standard library, pip install-ing it will cause errors

Ensure that the runtime for Python in the lambda is >= 3.7
Use an asyncio handler that will force the functionality to return results in a

synchronous manner

Forcing synchronous behavior

An example of the implementation below utilizes a run loop that asyncio provides to ensure we're gathering synchronous results.

import asyncio

loop = asyncio.get_event_loop()


# Some sample async function
async def my_async_func(event):
    await print(f"{event}")
    return 0


# Entrypoint for our lambda
def lambda_handler(event, context):

    # Call our async function
    coroutine = my_async_func(event)

    # Tell asyncio to run the function and gather the
    # results synchronously
    results = loop.run_until_complete(*[coroutine])

    return results

Another example that I haven't personally tested, but have seen in StackOverflow posts and whatnot:

import asyncio


# Some sample async function
async def my_async_func(event):
    await print(f"{event}")
    return 0


# Entrypoint for our lambda
def lambda_handler(event, context):

    # Call our async function
    results = asyncio.run(my_async_func(event))

    return results