AWS Lambda Stop Running EC2 Instances

An AWS Lambda that stops all running EC2 instances based on tags.
This Lambda could be run automatically by a CloudWatch Event (every day or so).
The main purpose of this lambda is to help you reduce your AWS bills.

How it works

This Lambda is based on EC2 tags to decide whateven an instance must be keep running or not.
It uses the RunMode tag that can take two values:

• OnDemand
  The instance is automatically stopped by the Lambda nad must be restarted manually.
• AlwaysOn
  The instance will not be stopped by the Lambda.

If an instance does not have the RunMode tag nor it has an unrecognized value, the Lambda assumes the OnDemand value.

IAM Role

This Lambda requires an execution role to work properly.
The IAM role must have the following policies:

• The AWSLambdaBasicExecutionRole AWS-managed policy.
• A customer-managed policy with these permissions:
  • ec2:DescribeInstances
  • ec2:StopInstances

Return values

This Lambda returns to different values:

• If no instance has been shutted down, the Lambda returns No OnDemand instance running.
• If one or many instances have been shutted down, the Lambda returns the list of the stopped instance's IDs.

These values are logged to CloudWatch by using the console.log function.

CloudFormation

A CloudFormation template is provided to quickly deploy all the needed resources.
As a prerequisite, you need to create a bucket which will contain the Lambda source code.

1. Once you bucket is created, package the application:

$ aws cloudformation package --template-file .\cloudformation.json --s3-bucket <bucket name> --output-template-file packaged.json

2. Then, deploy the CloudFormation stack:

$ aws cloudformation deploy --template-file .\packaged.json --stack-name ec2-autostop --capabilities CAPABILITY_IAM