Integration with S3

[jcarta]

One of my lambda functions makes calls to and from an S3 bucket. How would I go about testing this locally on my machine?

[Michael Hart]

docker run allows you to pass in env variables using -e: https://docs.docker.com/engine/reference/run/#env-environment-variables

So you can pass in AWS_ACCESS_KEY_ID and AWS_SECRET_ACCESS_KEY when you run the container.

[Michael Hart]

You can also share your local AWS config in the container – though this has lower fidelity with running a live Lambda (which uses env vars for credentials).

Eg:

docker run --rm \
  -v "$HOME"/.aws:/home/sbx_user1051/.aws:delegated,ro \
  -e AWS_PROFILE=myprofile \
  -e AWS_SDK_LOAD_CONFIG=1 \
  lambci/lambda:nodejs12.x

You can test whether this loads your credentials or not without invoking your Lambda:

docker run --rm \
  -v "$HOME"/.aws:/home/sbx_user1051/.aws:delegated,ro \
  -e AWS_PROFILE=myprofile \
  -e AWS_SDK_LOAD_CONFIG=1 \
  -w /var/runtime \
  --entrypoint node \
  lambci/lambda:nodejs12.x \
  -e 'new (require("aws-sdk").S3)().listBuckets().promise().then(console.log)'

[jcarta]

When you input your AWS credentials, is the actual (deployed) lambda function being run or is it still the local version?

[Michael Hart]

Still local. It's just allowing the lambda function that's running locally to use those credentials