Terraform script for defining stop and start functionality for EC2 instances

main.tf

resource "aws_iam_policy" "stop_start_ec2_policy" {
  name = "StopStartEC2Policy"
  path = "/"
  description = "IAM policy for stop and start EC2 from a lambda"

  policy = <<EOF
{
  "Version": "2012-10-17",
  "Statement": [
    {
      "Effect": "Allow",
      "Action": [
        "logs:CreateLogGroup",
        "logs:CreateLogStream",
        "logs:PutLogEvents"
      ],
      "Resource": "arn:aws:logs:*:*:*"
    },
    {
      "Effect": "Allow",
      "Action": [
        "ec2:Start*",
        "ec2:Stop*",
        "ec2:DescribeInstances*"
      ],
      "Resource": "*"
    }
  ]
}
EOF
}

resource "aws_iam_role" "stop_start_ec2_role" {
  name = "StopStartEC2Role"

  assume_role_policy = <<EOF
{
  "Version": "2012-10-17",
  "Statement": [
    {
      "Action": "sts:AssumeRole",
      "Principal": {
        "Service": "lambda.amazonaws.com"
      },
      "Effect": "Allow",
      "Sid": ""
    }
  ]
}
EOF
}

resource "aws_iam_role_policy_attachment" "lambda_role_policy" {
  role = "${aws_iam_role.stop_start_ec2_role.name}"
  policy_arn = "${aws_iam_policy.stop_start_ec2_policy.arn}"
}

resource "aws_lambda_function" "stop_ec2_lambda" {
  filename      = "ec2_lambda_handler.zip"
  function_name = "stopEC2Lambda"
  role          = "${aws_iam_role.stop_start_ec2_role.arn}"
  handler       = "ec2_lambda_handler.stop"

  source_code_hash = "${filebase64sha256("ec2_lambda_handler.zip")}"

  runtime = "python3.7"
  memory_size = "250"
  timeout = "60"
}

resource "aws_cloudwatch_event_rule" "ec2_stop_rule" {
  name        = "StopEC2Instances"
  description = "Stop EC2 nodes at 19:00 from Monday to friday"
  schedule_expression = "cron(0 19 ? * 2-6 *)"
}

resource "aws_cloudwatch_event_target" "ec2_stop_rule_target" {
  rule      = "${aws_cloudwatch_event_rule.ec2_stop_rule.name}"
  arn       = "${aws_lambda_function.stop_ec2_lambda.arn}"
}

resource "aws_lambda_permission" "allow_cloudwatch_stop" {
  statement_id  = "AllowExecutionFromCloudWatch"
  action        = "lambda:InvokeFunction"
  function_name = "${aws_lambda_function.stop_ec2_lambda.function_name}"
  principal     = "events.amazonaws.com"
}

resource "aws_lambda_function" "start_ec2_lambda" {
  filename      = "ec2_lambda_handler.zip"
  function_name = "startEC2Lambda"
  role          = "${aws_iam_role.stop_start_ec2_role.arn}"
  handler       = "ec2_lambda_handler.start"

  source_code_hash = "${filebase64sha256("ec2_lambda_handler.zip")}"

  runtime = "python3.7"
  memory_size = "250"
  timeout = "60"
}

resource "aws_cloudwatch_event_rule" "ec2_start_rule" {
  name        = "StartEC2Instances"
  description = "Start EC2 nodes at 6:30 from Monday to friday"
  schedule_expression = "cron(30 6 ? * 2-6 *)"
}

resource "aws_cloudwatch_event_target" "ec2_start_rule_target" {
  rule      = "${aws_cloudwatch_event_rule.ec2_start_rule.name}"
  arn       = "${aws_lambda_function.start_ec2_lambda.arn}"
}

resource "aws_lambda_permission" "allow_cloudwatch_start" {
  statement_id  = "AllowExecutionFromCloudWatch"
  action        = "lambda:InvokeFunction"
  function_name = "${aws_lambda_function.start_ec2_lambda.function_name}"
  principal     = "events.amazonaws.com"
}

Can you also provide the steps to run this? I have a python file with stop and start code, I have downloaded terraform on my machine.. Can you please share the steps to stop or start any instance, please?

You need to create a terraform project first and move the above code to main.tf and add also the provider and var files
commands to apply to terraform

terraform init -backend-config="${Your_vars_file}"
terraform plan -var-file="${Your_vars_file}" -lock=false
terraform apply

commands to run the above code

# config
terraform init -backend-config="${Your_vars_file}"; 
terraform init -var-file="${Your_vars_file}"; 
zip ec2_lambda_handler.zip ec2_lambda_handler.py

#apply
terraform apply -var-file="${Your_vars_file}"

#destory
terraform destroy -var-file="${Your_vars_file}"

Thank you so much for providing this. I am new to terraform can you please share your variables.tf?

it could be something like

variable "region" {
  type = "string"
  description = "Create an ec2 instacne in a given region"
}
variable "ami" {
  type = "string"
  description = "Image ami id"
}
variable "instance_type" {
  type = "string"
  description = "Instance type"
}
variable "instance_name" {
  type = "string"
  description = "Name to be used on all resources as prefix"
}
variable "name" {
  type = "string"
  description = "Name of all the resources"
}
variable "instance_count" {
  description = "Number of instances to launch"
  type        = number
  default     = 1
}
variable "profile" {
  default     = "terraform"
  description = "Using terraform for deployment"
}

and you may need another file for the values of these vars

region="eu-central-1"
ami="ami-dd3c0f36"
instance_type="t2.xlarge"
instance_name="dev01-docker01"
instance_count=1
instance_size="100"

You are a life saver!! We are implementing it in our dev environment. Let say if I would like to use different tags based on the region to make more functional. Obviously I would have to add the necessary code to the main.tf by updating the time but I am scratching my head on the lambda part a I am new to it. I would appreciate any feedback you have hero. See my code below :

import boto3
region = 'us-gov-west-1'
ec2 = boto3.client('ec2', region_name=region)
response = ec2.describe_instances(Filters=[
        {
            'Name': 'tag:Auto-Start-EST',
            'Values': [
                'true',
            'Name': 'tag:Auto-Start-MST',
            'Values': [
                'true',
            'Name': 'tag:Auto-Start-PST',
            'Values': [
                'true',
            ]
        },
    ])

instances = []

for reservation in response["Reservations"]:
    for instance in reservation["Instances"]:
        instances.append(instance["InstanceId"])

def stop(event, context):
    ec2.stop_instances(InstanceIds=instances)
    print('stopped instances: ' + str(instances))

def start(event, context):
    ec2.start_instances(InstanceIds=instances)
    print('started  instances: ' + str(instances))

Never mind my question. To make things simpler I will just create a module for each region (EST, PST, MST) with their respective variables and output tfs. Thoughts!!! ;)

it seems that you have a syntax issue here, you are missing couple of ]

response = ec2.describe_instances(Filters=[
        {
            'Name': 'tag:Auto-Start-EST',
            'Values': [
                'true',
            'Name': 'tag:Auto-Start-MST',
            'Values': [
                'true',
            'Name': 'tag:Auto-Start-PST',
            'Values': [
                'true',
            ]
        },
    ])

Can you please provide for the Azure too to Start and stop the VM's