Schedule your EC2 instances to stop and start using Lambda(boto3) and Cloudwatch

Just take an example that you have a very large organization over 100 people and your instances are not being used after business hours but you are still paying for that unused time. It would be a very tiresome process to stop all the instances manually if it is spread across all the regions.

We can automate by following process:

1. Create a Lambda Function with Python 3.7( I used this python version while implementing in my organization. Yours may have got updated but the code will be following python 3.7 rules. So follow accordingly),
2. For roles, first create a policy in IAM by pasting the following code in JSON:

{
 “Version”: “2012–10–17”,
 “Statement”: [
 {
 “Effect”: “Allow”,
 “Action”: [
 “logs:CreateLogGroup”,
 “logs:CreateLogStream”,
 “logs:PutLogEvents”
 ],
 “Resource”: “arn:aws:logs:*:*:*”
 },
 {
 “Effect”: “Allow”,
 “Action”: [
 “ec2:DescribeInstances”,
 “ec2:DescribeRegions”,
 “ec2:StartInstances”,
 “ec2:StopInstances”
 ],
 “Resource”: “*”
 }
 ]
}

Select this policy while creating Lambda Role. Select this role while creating your Lambda function. This roles allow the function to access Cloudwatch logs, events and EC2 instance( stop-start and to describe region in which the instance is present)