EC2 Instances Scheduling

This doc explains about the scheduling of aws ec2 instances. Below are the explanation.

EC2

In this we have setup scheduling for common ai central training servers.

Setup scheduling for common ai instances in which instances are being stopped at 9PM and started at 10AM.
Scheduling is setup from monday to friday only. and on weekend servers will be totally stopped.
We have setup lambda function which is triggered by Amazon EventBridge scheduler.
Lambda function
- centralised_ai_instances_schedule
- In this function, We have created aliases based on env and operation.
- Now, there are 4 aliases for development and production as mentioned below.
  - For development: scheduleON-dev, scheduleOFF-dev
  - For production: scheduleOFF, scheduleON
Created 2 Rules in Amazon event bridge for on and off operation based on lambda function alias. below are the rules
- central_ai_instances_scheduler_on
- central_ai_instances_scheduler_off
Jenkins job:
- There is 4 jenkins jobs for code deployments on lambda function based on operation like on & off for development & production. Below are the jobs link.
- Jobs
- Instances configurations are fetching from s3 in jobs.
- We store instances configurations like instanceid, instances ontime, offtime on s3 in below location.
  - s3://oriserve-config/Tools/centralised_ai_instance_scheduler/Lambda/Environment/Development/
  - s3://oriserve-config/Tools/centralised_ai_instance_scheduler/Lambda/Environment/Production/
- Below are the file format

    {
        "instanceid":"instanceID",
        "ontime": "00:00",
        "offtime": "00:00"
    }

If there is a req for adding/updating servers for scheduling then first download the file from s3 from development folder then add/update your instances details in above format and upload to s3.
After tested in development then update the same for production with approval of devops lead.