Serverless SSM Fetch

Serverless SSM Fetch is an "AWS provider only" plugin that allows to fetch parameters from AWS Store Parameters and assign them to serverless.yml functions environment variables.

Before using this plugin you must have set your parameters into AWS System Manager Parameter Store

NOTE: to use this plugin, the AWS credentials that you use for your project must have permissions for:

  • ssm:Describe*
  • ssm:Get*
  • ssm:List*

Versions

  • For serverless framework v2: use serverless-ssm-fetch@1.x
  • For serverless framework v3: use serverless-ssm-fetch@2.x

Setup

First, you have to set serverless-ssm-fetch plugin by running:

serverless plugin install --name serverless-ssm-fetch

This will install the required npm package and add the plugin to your serverless.yml file.

...
plugins:
- serverless-ssm-fetch
...

Then, you must declare the SSM Parameters that must be assigned to your functions environment variables.

It consists in key value pairs, where the key is the environment variable name you want to use, and the value is the parameter path you set in AWS System Manager Parameter Store on your AWS account. Example: APP_ID: /aws/ssm/parameter/path/app_id.

To declare them, use the serverlessSsmFetch accessor inside custom variable in your serverless.yml file. FYI, it also works if you use a nested file for your custom (custom: $(file:./path/to/file)).

...
custom:
serverlessSsmFetch:
APP_ID: /aws/ssm/parameter/path/app_id
APP_KEY: /aws/ssm/parameter/path/app_key
APP_SECRET: /aws/ssm/parameter/path/app_secret~true
...

Usage

By default, there is nothing more to do if you want all your SSM Parameters injected in all your functions. But of course, you will probably want to assign specific parameters to specific functions. You can specify it on a per function basis this way:

...
custom:
serverlessSsmFetch:
APP_ID: /aws/ssm/parameter/path/app_id
APP_KEY: /aws/ssm/parameter/path/app_key
APP_SECRET: /aws/ssm/parameter/path/app_secret~true
functions:
hello:
handler: handler.hello
ssmToEnvironment:
- APP_ID
environment:
- NODE_ENV: development
...

This will add ONLY the SSM Parameter APP_ID to the function environment variables. In this case, the function hello will be pushed to AWS Lambda with environment variables NODE_ENV and APP_ID.

Decryption

On AWS Parameters Store you can decide to encrypt parameters when you set them. Meaning that you must decrypt them with your KMS key when you use them within your application.

With serverless-ssm-fetch you can decide to decrypt them at runtime to add the decrypted format of the parameter to your environment variable. It is useful if you don't use an AWS KMS key to decrypt parameters within your app.

To decrypt a parameter you just have to add ~true at the end of your AWS SSM Parameter path:

...
custom:
serverlessSsmFetch:
APP_SECRET: /aws/ssm/parameter/path/app_secret~true
...

Licensing

MIT License - Copyright © 2017 Emmanuel Lemoine