Skip to content

Latest commit

 

History

History
82 lines (51 loc) · 4.41 KB

README.md

File metadata and controls

82 lines (51 loc) · 4.41 KB
Raw

AWS Lambda function to build and assess GitHub projects in SWAMP (unoffical)

This is an attempt to have a SWAMP integration for GitHub project (like Travis CI and all of these: https://github.com/integrations). This is accomplished using AWS Lambda

This is not an official SWAMP project

It is done this way:

  1. For a github project, setup an Amazon's Simple Notification Service (AWS SNS) service for events (like push). GitHub services are pre-built integrations that perform certain actions when events occur on GitHub. For details on services, go to your GitHub project home page (you must be an admin for the project) >> settings >> integrations & services.

  2. When an event like (git push) happens for the registered GitHub project, GitHub sends a event notification using AWS SNS to a registered end point, in this case it is the AWS Lambda function github-to-swamp (this project).

  3. The github-to-swamp lambda function parses the event object, downloads the latest archive of the master branch of the github project which sent the notification.

  4. The github-to-swamp lambda function uses the swamp-api to upload the archive with build information (provided in package.conf configuration file) to SWAMP. The github-to-swamp lambda function then triggers assessments with all the relevant static analysis tools in SWAMP. Users will get a notification from SWAMP once the assessments are complete.

AWS Lambda and GitHub Webhooks setup

To use this lambda function, you first have to setup an AWS account:

  1. Open an AWS account (if don't already have one) and an AWS IAM user (http://docs.aws.amazon.com/lambda/latest/dg/setting-up.html)
  2. Create an AWS SNS topic (http://docs.aws.amazon.com/sns/latest/dg/GettingStarted.html)
  3. Create an AWS Lambda function (https://aws.amazon.com/lambda/getting-started/) that uses python-2.7 runtime.

This blog has step-by-step tutorial to accomplish the above:

https://aws.amazon.com/blogs/compute/dynamic-github-actions-with-aws-lambda/

Usage

First clone or download this project

Make

To make a zip file that contains the lambda function and its dependencies. First, fill in resources/user-info.conf with your SWAMP login credentials and run ./scripts/make.sh, if everything is fine, this will create release/github-to-swamp.zip.

This lambda function has two external dependencies that it carries along with it.

  1. Requests: HTTP for Humans
  2. SWAMP Python API (unofficial)

Upload

Users can upload the zip file using AWS lambda web interface or running the command scripts/upload.sh.

The script uses AWS CLI. You can install AWS CLI by running pip install awscli.

The scripts/upload.shneeds the AWS lambda function name and AWS configuration profile name as argument. There is an optional third argument for AWS region which defaults to us-west-2.

Package.conf in GitHub Project

Add a file named package.conf in the top-level directory of your GitHub project. The contents of the package.conf file must look like this

package-short-name=my-java-pkg
package-version=<filed-in-by-lambda-function>
package-archive=<filed-in-by-lambda-function>
package-dir=<filed-in-by-lambda-function>
package-language=Java
# supported ant, ant+ivy, maven, gradle
build-sys=maven
build-file=pom.xml
build-target=clean install
build-opt=-DskipTests -Dmaven.test.skip
config-cmd=mvn
config-opt=clean
config-dir=.

Test

Users can test/trigger the AWS lambda function using AWS web interface or the command line using ./scripts/invoke.sh.

The script used default GitHub event resources/sample_github_sns_event.json and requires ARN for the AWS Lambda function as the argument.

Limitations

  1. The AWS Lambda runtime environment does not have git installed. Because of this the actual commit reference that generated the event cannot be fetched. i.e. the lambda function always downloads the archive of the latest of the master branch.

  2. Currently, only Java package upload and assessments are possible, and this is because of the limitations of SWAMP Python API (unofficial)

  3. Requires build information be specified in a package.conf in the top-level directory of the GitHub project.