henrist/sceptre-s3-packager

S3 packager for Sceptre 2

sceptre-s3-packager

Hook and resolver for Sceptre v2 to package code dependencies and uploading it to an S3 bucket for usage in a template.

Inspired by https://github.com/cloudreach/sceptre-zip-code-s3

Getting started

Only Python 3 is actively being supported.

Install using pip:

pip install sceptre-s3-packager

Use the hook and resolver in your template config:

template_path: ...
hooks:
  before_create:
    - !sceptre_s3_upload ./directory-to-zip-to-s3
  before_update:
    - !sceptre_s3_upload ./directory-to-zip-to-s3
sceptre_user_data:
  Code:
    S3Bucket: my-s3-bucket
    S3Key: !sceptre_s3_key ./directory-to-zip-to-s3

Use the data in the template, e.g. by using Jinja2 template with something like:

  MyLambda:
    Type: AWS::Lambda::Function
    Properties:
      Code:
        S3Bucket: {{ sceptre_user_data.Code.S3Bucket }}
        S3Key: {{ sceptre_user_data.Code.S3Key }}

Usage

Hook:

• !sceptre_s3_upload <directory> (reads S3Bucket from sceptre_user_data.Code.S3Bucket)
• !sceptre_s3_upload <directory>^^<s3-bucket>

Resolver:

• !sceptre_s3_key <directory> (returns a path where the packaged directory is uploaded, e.g. sceptre/68063a99bb6d95401d688d28f19ee412)

Details

The hook will zip the contents of the directory and upload it as sceptre/MD5HASH to the S3-bucket, with MD5HASH being md5 hash of the zip file content.

When zipping all files will be given a fixed modification time, so that only the contents of the files are used to determine the upload file and cause invalidation on changes.