arc-plugin-lambda-env  

Arc serverless framework (arc.codes) plugin for adding env variables to your lambdas from architect created resources

This plugin will add environment variables with logical names of resources created by your .arc schema.

Initial support only for @tables (dynamodb table names)

Install

npm i --save-dev arc-plugin-lambda-env

Usage

After installing add @plugins and @lambda-env-vars pragmas to your app.arc file:

app.arc

@app
myapp

@http
get /

@tables
mytable
  pk *String
  sk **String
tabletwo
  pk *String
  sk **String

@lambda-env-vars
tables
  mytable MY_APP_TABLE_NAME
  tabletwo TABLE_TWO_NAME

@plugins
arc-plugin-lambda-env

Options

This plugin supports the following options under the @lambda-env-vars pragma:

|Option|Description|Example| |---|---|---| |tables| A list of tablename => ENV_VAR_NAME to be added to every lambda in your app|tables  tablename ENV_VAR_NAME|

Sandbox

Running arc sandbox will run your lambdas locally with the correct env variables defined. In order to In sandbox, you'll need to write code to determine what the local dynamodb endpoint is. See an example in the sample-app code. Arc 10 adds a ARC_SANDBOX ENV variable containing a JSON string that can be parsed to find the locally running port in ports.tables. In deployed environments ARC_SANDBOX will be undefined and your clients should automatically connect to the right endpoint/region based on aws lambda injected vars.

Sample Application

There is a sample application located under sample-app/. cd into that directory, npm install and you can run locally via npm start or npx arc sandbox or deploy to the internet via npx arc deploy.

⚠️ Known Issues

Supported architect versions

This plugin will work for architect version 10.0.0-RC.2 and above. Please use version 0.1.0 of this plugin if you have architect version 9.0.2 or below. Any versions in between are currently not supported

Only @tables (dynamodb) supported

I'll be adding support for injecting other resources identifiers as environment variables to your lambdas like @events (sns) and @queues (sqs).

Can't target specific lambdas

As of now, declared env variables go to all lambda functions. Would be nice to specify which lambdas get which env variables added.

Other Ideas

Config file vs ENV Vars

It might be nicer to write a config file, probably in the universal json, on build/deploy that can be imported in specified lambdas rather than using environment variables.