AWS EC2 instance backup using AMI (pyton/boto3)

Alpha ALERT !

This is totaly work in progress ... not suitable for usage jet

How does it work

Start it from cron and hope for the best :)

Workflow

1. get all instances with tag backup set to true
2. for each of above instance

• create AMI
• tag AMI with tags needed for recovery (ip, subnet, name ...)
• check instance retention tag
• if there is a AMI that is older then number of days defined by retention and number of backups (AMIs) for this instance is grater than retntion
  1. deregister it
  2. delete all snapshots belonging to this AMI

Used tags

Instance

• backup (true | false) - if set to true instance will be backuped
• retention - define backup retention. If not set retention from config file will be used

AMI

• Name - ami name in format BACKUP +instance_name+date
• created - in time() format (epoch)
• srcInstanceId - source instance id
• srcInstanceName - source instance Name tag
• srcPrimaryIP - source instance primary IP (for now)
• srcSubnetId - source instance subnet id
• srcSecurityGroupId - source instance first (for now) security group
• srcInstanceType - source instance type (m4.large, t2.medium, ...)

sample config file (config.yaml)

defaults:
  retention: 7
  log file path: "./brokenSpectre.log"
  # [10,20,30,40,50]... higher number, less logs (https://docs.python.org/2/library/logging.html#logging-levels)
  log level: 10
  boto log level: 50
  delete abandoned: False
aws account id: "1234567890"