Simple framework to develop and deploy AWS Lambda functions written in Python.
- working setup of
make(should be on every modern Linux/UNIX box) - working setup of
pip(either with or without virtualenv) - working setup of the AWS CLI tool (including appropriate credentials)
Edit the provided Makefile and add the ARN of an approprite IAM Role. This
IAM Role has to exist in your AWS account and should allow at least
logs:CreateLogStream and logs:PutLogEvents as a basic policy. Details
on setting up an appropriate IAM Role for you Lambda function can be found in the
AWS Lambda Permissions Model
documentation.
Use the skeleton provided by lambda_function.py to get started. AWS Lambda
will call the function lambda_handler inside by default.
Add required Python modules to LIBS in the Makefile and run
make deps. Dependencies of the specified LIBS will not be resolved
automatically: To install LIBS make will call pip install with
--no-deps to be able to cleanup properly with make clean. You might
not need to vendor in every module dependency since AWS Lambda already provides
most of the modules. In question, or if you want to use a specific version, add
it to the LIBS in the Makefile.
Use make deploy in the console to deploy your current code to AWS Lambda.
This will ensure your AWS Lambda function is created with or updated to your
current settings specified in the Makefile, package your code, including
the LIBS specified in the Makefile and deploy it to your AWS Lambda
function.
Use make run to deploy and run your application on AWS Lambda. This will
do everything included in make deploy plus invoking the AWS Lambda
function.
Use make invoke to run the currently deployed version of your AWS Lambda
function. make invoke is not only faster if you just want to rerun your
code, but also gives you some parsable output if you return JSON as a response
for example.
Use make the console to see more available options.