aws-amicleaner

Cleanup your old unused ami and related snapshots

|Travis CI| |codecov.io| |pypi|

Maintenance note

This project is not under active maintenance but still active. Consider using the official AWS-supported Amazon Data Lifecycle Manager <https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/snapshot-lifecycle.html>__.

Description

This tool enables you to clean your custom Amazon Machine Images (AMI) <https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/AMIs.html>__ and related EBS Snapshots <https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/EBSSnapshots.html>__.

You can either run in fetch and clean mode where the tool will retrieve all your private AMIs and EC2 instances, exclude AMIs being holded by your EC2 instances (it can be useful if you use autoscaling, and so on ...). It applies a filter based on their names or tags and a number of previous AMIs you want to keep. You can also check and delete EBS snapshots left orphaned by manual deletion of AMIs.

It can simply remove AMIs with a list of provided ids.

Prerequisites

• awscli <http://docs.aws.amazon.com/cli/latest/userguide/installing.html>__
• python 2.7 or 3+
• python pip <https://pip.pypa.io/en/stable/installing/>__

This tool assumes your AWS credentials are in your environment, either with AWS credentials variables :

.. code:: bash

export AWS_DEFAULT_REGION='your region'
export AWS_ACCESS_KEY_ID='with token Access ID'
export AWS_SECRET_ACCESS_KEY='with token AWS Secret'

or with awscli :

.. code:: bash

export AWS_PROFILE=profile-name

Minimum AWS IAM permissions

To run the script properly, your ``aws`` user must have at least these
permissions in ``iam``:

.. code:: json

    {
        "Version": "2012-10-17",
        "Statement": [
            {
                "Sid": "Stmt1458638250000",
                "Effect": "Allow",
                "Action": [
                    "ec2:DeleteSnapshot",
                    "ec2:DeregisterImage",
                    "ec2:DescribeImages",
                    "ec2:DescribeInstances",
                    "ec2:DescribeSnapshots",
                    "autoscaling:DescribeAutoScalingGroups",
                    "autoscaling:DescribeLaunchConfigurations"
                ],
                "Resource": [
                    "*"
                ]
            }
        ]
    }

Installation
------------

amicleaner is available on pypi and can be installed on your system with pip

From pypi
~~~~~~~~~

.. code:: bash

    [sudo] pip install aws-amicleaner

From source
~~~~~~~~~~~

You can also clone or download from github the source and install with pip

.. code:: bash

    cd aws-amicleaner/
    pip install [--user] -e .

Usage
-----


Getting help
~~~~~~~~~~~~

.. code:: bash

    amicleaner --help


Fetch and clean
~~~~~~~~~~~~~~~

Print report of groups and amis to be cleaned

.. code:: bash

    amicleaner --full-report

Keep previous number of AMIs

.. code:: bash

    amicleaner --full-report --keep-previous 10

Regroup by name or tags

.. code:: bash

    amicleaner --mapping-key tags --mapping-values role env

Exclude amis based on tag values

.. code:: bash

    amicleaner --mapping-key tags --mapping-values role env -excluded-mapping-values prod

Skip confirmation, can be useful for automation

.. code:: bash

    amicleaner -f --keep-previous 2


Activate orphan snapshots checking

.. code:: bash

amicleaner --check-orphans

Delete a list of AMIs

.. code:: bash

    amicleaner --from-ids ami-abcdef01 ami-abcdef02


.. |Travis CI| image:: https://travis-ci.org/bonclay7/aws-amicleaner.svg?branch=master
   :target: https://travis-ci.org/bonclay7/aws-amicleaner
.. |codecov.io| image:: https://codecov.io/github/bonclay7/aws-amicleaner/coverage.svg?branch=master
   :target: https://codecov.io/github/bonclay7/aws-amicleaner?branch=master
.. |pypi| image:: https://img.shields.io/pypi/v/aws-amicleaner.svg
   :target: https://pypi.python.org/pypi/aws-amicleaner


See this `blog article
<https://blog.d2si.io/2017/06/15/garbage-collecting-amazon-machine-images-ami/>`__
for more information.