Skip to content

Latest commit

 

History

History
143 lines (99 loc) · 4.86 KB

README.rst

File metadata and controls

143 lines (99 loc) · 4.86 KB

Lambda-httpx

Lambda-httpx; use familiar async httpx library to access HTTP enabled (simple proxy) AWS Lambda functions

Quick start

Installation

pip intall lambda-httpx

Usage Examples

Lambda-httpx provides a transport which can be mounted with the httpx client. This then relays over all requests with http+lambda:// to AWS lambda whose name matches the hostname.

Using from existing event loop:

import httpx
from lambda_httpx import AsyncLambdaTransport
# ...
async with AsyncLambdaTransport() as transport:
    mounts = {"http+lambda://": transport}
    async with httpx.AsyncClient(mounts=mounts) as client:
        response = await client.get("http+lambda://flaskexp-test/health")

Stand alone that calls endpoint 10 times asyncronously.

import asyncio
import httpx
from lambda_httpx import AsyncLambdaTransport

async def main(count):
    async with AsyncLambdaTransport() as transport:
        mounts = {"http+lambda://": transport}
        async with httpx.AsyncClient(mounts=mounts) as client:
            coros = [client.get("http+lambda://flaskexp-test/health") for _ in range(count)]
            return await asyncio.gather(*coros)

if __name__ == "__main__":
    loop = asyncio.get_event_loop()
    responses = loop.run_until_complete(main(10))
    print(responses)

A synchronous Transport also exists and can be utilized as follows:

import httpx
from lambda_httpx import LambdaTransport

with LambdaTransport() as transport:
    mounts = {"http+lambda://": transport}
    with httpx.Client(mounts=mounts) as client:
        response = client.get("http+lambda://flaskexp-test/health")
        print(response)

Lambda authorization is configured via boto3, and can be set up using environment variables or a configuration file. Configuration file is recommended. Example credential file ~/.aws/credentials:

[default]
aws_access_key_id =  XXXXXXXXXXXXXXXXXXXX
aws_secret_access_key = XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

Similar to authorization, region can be configured either via the environment variable AWS_DEFAULT_REGION, configuration file. Region can also be set on initialization of AsyncLambdaTransport(region="us-west-2"). Example configuration file ~/.aws/config:

[profile default]
region = us-west-2

The lambdas must support proxy integration, which is used commonly by frameworks such as Zappa, Mangum.

Why

In using REST microservice architecture it is important to be able to conveniently make calls from one service to another. To use this pattern in AWS serverless ecosphere along with Lambda one is practically forced to stand up an API Gateway in front of the lambda. This has several distinct disadvantages, all mostly along the lines of security.

  • API Gateway publicly exposes endpoints
  • API Gateway uses own authentication / authorization schema. While Lambda already supplies us with IAM.
  • Extra dependencies in call chain. While availability is high, latency may still be of concern.

Over all, to reduce exposure of private sub-services, re-use IAM authentication / authorization and reduce latency.

How does its work

Simple, we register a scheme name with httpx and use a lambda specific transport adapter which translates a httpx request to lambda invoke compatible with AWS API Gateway simple proxy format.

See also

  • Lambda-requests: Similar library that allows same functionality via python requests library.