Skip to content

The repository for the rearchitected site-scanning project, specifically the scanning engine itself.

Notifications You must be signed in to change notification settings

GSA/site-scanning-engine

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Site Scanning Engine

CodeQL Semgrep Snyk Scan MegaLinter woke

Description

This repository is for the Site Scanning engine itself, the codebase that actually runs the scans and generates data. This is the new base scanner repository which uses Headless Chrome, powered by Puppeteer for scanning.

For more detailed documentation about the Site Scanning program, including who it's for, what it does, long-term goals, etc. please visit the Site Scanning program website, especially the Technical Details page.

The project's issue tracker and other relevant repositories and links can be found here.

Table of Contents

Quickstart

Development Requirements:

  • git
  • nodejs
  • nvm (see .nvmrc for current node version.
  • docker
  • docker-compose
  • Cloud Foundry CLI (aka cf)
  • redis-cli (optional)

Installation

First clone the repository:

git clone https://github.com/GSA/site-scanning-engine/

From the project root run:

nvm use

This will install the correct Node version for the project.

npm i

This will install all production and development Node dependencies.

Dotenv

The project uses a dotenv (.env) file for local development credentials. Note that this file is not version-controlled and should only be used for local development.

Before starting Docker, create a .env file in the project root and add the following values replacing <add_a_key_here> with a local passwords that are at least 8 characters long.

Note: this is only for local development and has no impact on the Cloud.gov configuration

# postgres configuration
DATABASE_HOST=localhost
DATABASE_PORT=5432
POSTGRES_USER=postgres
POSTGRES_PASSWORD=<add_a_key_here>

# redis configuration
QUEUE_HOST=localhost
QUEUE_PORT=6379

# Minio Config -- Minio is an S3 api compliant storage
MINIO_ACCESS_KEY=<add_a_key_here>
MINIO_SECRET_KEY=<add_a_key_here>
AWS_ACCESS_KEY_ID=<add_a_key_here>
AWS_SECRET_ACCESS_KEY=<add_a_key_here>
S3_HOSTNAME=localhost
S3_PORT=9000
S3_BUCKET_NAME=site-scanning-snapshot

# Sets the development environment name to dev
NODE_ENV=dev

Docker

From the project root run:

docker-compose up --build -d

This will build (--build) all of the Docker containers and network interfaces listed in the docker-compose.yml file and start them running in the background (-d).

docker-compose down will stop and remove all containers and network interfaces.

Running docker-compose up --build -d will rebuild all of the containers. This is useful if you need to wipe data from the database, for instance.

If you encounter any issues starting the containers with docker-compose, specifically related to OOM errors (or Exit 137) try upping the resources in your Docker preferences.

Building application images

To build the application image, go to the project root and run:

docker build -f apps/scan-engine/Dockerfile .

Build and Start all apps

cd to the project root and run:

npm run build:all

This command will build the apps, which compiles from Typescript to Javascript, doing any minification and optimization in the process. All of the app artifacts end up in the /dist directory. This is ultimately what gets pushed to Cloud Foundry.

Note that you can also build apps seperately:

npm run build:api
npm run build:scan-engine
npm run build:cli

Next, you can start the apps with following command:

npm run start:all

The apps are started as follows: first the API starts and then the Site Scanning worker follows. This is designed so that the API app runs any shared configuration against the database first.

Note, that you can start the apps individually as follows:

npm run start:api
npm run start:scan-engine

Ingest Website List

The Site Scanning engine relies on a list of federal domains and metadata about those to domains to operate. This list is ingested into the system from a public repository using a the Ingest Service.

To run the ingest service do the following:

npm run ingest -- --limit 200

The limit parameter is optional, but it can be useful to use a smaller subset of the total list for local development.

Enqueue scans

To enqueue for scan all sites in the website table:

npx nest start cli -- enqueue-scans

Run individual Scans

To scan a single site, which must be in the website table, run commands like:

npx nest start cli -- scan-site --url 18f.gov

NOTE: This is intended for testing scan behavior, and doesn't currently write results to the database.

Test

From the project root run:

npm run test:unit

This runs all unit tests.

Deploy

First, log in to Cloud.gov using the CLI and choose the organization and space.

Then, you can use the cloudgov-deploy.sh script to build and deploy the apps.

You can optionally pass a different manifest file with cloudgov-deploy.sh manifest-dev.yml.

Additional Documentation

About

The repository for the rearchitected site-scanning project, specifically the scanning engine itself.

Resources

Code of conduct

Security policy

Stars

Watchers

Forks