Skip to content

Latest commit

 

History

History
33 lines (23 loc) · 5.19 KB

infrastructure.md

File metadata and controls

33 lines (23 loc) · 5.19 KB
Raw

Infrastructure 🧱

The goal for all infrastructure in this project is to be as simple as possible while being robust, distributed, and scalable. We also aim to have all components open soruce, maintainable, and easy for contributors to contribute to.

Architecture Diagram 🏗️

To help maintainers and future contributors wrap their heads around the infrastructure we have deployed, we have created an architecture diagram which lays everything out.

diagram

Below is a description of the numbered labels in the diagram

  1. Scanners - The scanners are responsible for scanning data from the Escape from Tarkov game. This is the only portion of the project that is closed source. The scanners collect data and publish it to the main sql database for storage
  2. The Database - The database is where all flea market and other data is stored. The database is hosted in planetscale
  3. Tarkov Data Manager Cron Jobs - All cron jobs to sync the database with the API (cloudflare workers) are run through the tarkov-data-manager (TDM). The primary role of the TDM is for core contributors to update the database with new data, to check on scanners, or to update the schedule at which the cron jobs run. The cron jobs are very important as they keep the API fresh with data from our primary SQL database
  4. API - The API which powers everything for tarkov.dev is run using cloudflare workers - tarkov-data-api
  5. API Consumers - API consumers are anything and anyone that is using the GraphQL API. This could be you or one of the many projects using our API 😄. You can try out the API with the following playground - api.tarkov.dev/___graphql - A notable consumer of the GraphQL API is the Stash Discord Bot which is a project under this organization and hosted in Heroku
  6. tarkov.dev - The main website (tarkov.dev) is powered by Cloudflare Pages and uses the graphql API to render data for item prices, weapon stats, and more! - tarkov.dev site repo
  7. Tarkov Data Manager - The tarkov-data-manager is a tool used by primary contributors to this project. It manages data in the database, helps control the game scanners, and interacts with the cloudflare workers as well. It is hosted in an Azure VM - tarkov-data-manager
  8. Socket Server - The socket server is used with tarkov.dev to control the site with another device (mobile control feature). It is also used to collect console output from scanners with the the tarkov-data-manager. It is hosted in Heroku - tarkov-socket-server
  9. Static S3 Bucket - Static assets (mainly images) for tarkov.dev are stored in an S3 bucket. This bucket is used with Cloudflare's CDN
  10. Cloudflare CDN - The Cloudflare CDN is used to serve static assets (mainly images) to tarkov.dev site and the tarkov-data-manager. It reads these images from an S3 bucket and caches them on Cloudflare's edge to serve them quickly to clients in a distributed manner
  11. status.tarkov.dev - The status.tarkov.dev service is used to monitor the health for many of the components of the infrastructure. It is hosted on an Azure VM and uses GET healthchecks and PUSH based healthcheck to collect status metrics about other services

Open Source Notice 💡

The overwhelming majority of The Hideout's infrastructure is open source. That being said, there are a couple of components that are not open source at this time. Those items are listed below with a brief reason:

  • The Database - This database is hosted in planetscale and we do not have a good mechanism to open source it at this time since most of the work (schema, etc) requires access to the planetscale console and a login. It is restricted to trusted core contributors.
  • CDN - The CDN and its components (S3, Cloudflare, etc) are not open source at this time mainly because we do not have a good way to share them. They require access to AWS or the Cloudflare console and that is restricted to trusted core contributors. However, a large amount of our cloudflare configurations are open source as they are managed in a public repository with Terraform
  • Scanners - The scanners are closed source and for a good reason. Their code could be taken and modified to interact with the flea market to create trading bots. Our scanners simply view the flea market data and do not buy or sell items and never will. However, if the scanners were open source and modified they could do exactly that. We don't need any more flea market bots out there so these will be closed source for that reason.