Optimize EC2 cost with Lambda > Create Lambda Function > Function start instance

Function start instance

Create Lambda Function that executes Start instances

In the Lambda Function interface

Select Create function

Lambda Function start instance

In the Create function interface

Select Author from scratch
Function name, enter dc-common-lambda-auto-start
Runtime, select Python 3.8
Select x86_64

Lambda Function start instance

In the Create function interface

Select Change default execution role
Select Use an existing role
Select dc-common-lambda-role
Select Create function

Lambda Function start instance

Before importing the code, create Environment variables as follows:

Select Configuration
Select Environment variables
Select Edit

Lambda Function start instance

In the Edit environment variables interface

Key, enter environment_auto
Value, enter true
Select Save

Lambda Function start instance

After creating the environment

Select Code
Import source code: You need to change the webhook_url to receive notifications to Slack.

import boto3
import os
import json
import urllib3

ec2_resource = boto3.resource('ec2')
http = urllib3.PoolManager()
webhook_url = "https://hooks.slack.com/services/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"

def lambda_handler(event, context):
    environment_auto= os.environ.get('environment_auto')
    if not environment_auto:
        print('Target is empty')
    else:
        instances = ec2_resource.instances.filter(
            Filters=[{'Name': 'tag:environment_auto', 'Values': [environment_auto]}]
        )     
        if not list(instances):
            response = {
                "statusCode":500,
                "body": "Target Instance is None"
            }
        else:
            action_start = instances.start()
            sent_slack(action_start)
            response = {
                "statusCode":200,
                "body": "EC2 Starting"
            }
        return response

def sent_slack(action_start):
    list_instance_id = []
    if (len(action_start)>0) and ("StartingInstances" in action_start[0]) and (len(action_start[0]["StartingInstances"])>0):
        for i in action_start[0]["StartingInstances"] :
            list_instance_id.append(i["InstanceId"])
            msg = "Starting Instances ID:\n %s" %(list_instance_id)
            data = {"text":msg}
            r = http.request("POST",
                webhook_url,
                body = json.dumps(data),
                headers = {"Content-Type":"application/json"})
    else:
        print ('Not found Instances Start')

Save and select Deploy

Lambda Function start instance

Next create a CloudWatch rule

In the rule details section, Name, enter dc-common-lambda-auto-start
Description, enter dc-common-lambda-auto-start
Select Schedule
Select Next

Lambda Function start instance

In the Schedule pattern interface

There are 2 Schedule patterns to choose from:

A detailed schedule that runs at a specific time, such as 8am PST on the first Monday of each month.
Schedule runs at a regular pace, such as every 10 minutes

Select A schedule that runs at a regular rate, such as every 10 minutes.
Set rate, value is 8 hours
Select Next

Lambda Function start instance

In the Target interface

Select AWS service
Select Lambda function
Function, select dc-common-lambda-auto-start
Select Next

Lambda Function start instance

Select Next

Lambda Function start instance

Check again and select Create rule

Lambda Function start instance