Rewriting Services to Support Andromeda

[zllovesuki]

Andromeda/Whirlpool/Messier is the new infrastructure services to support geographically redundant operations. It can withstand up to N/2-1 datacenter failures and still remain operational. Similar to AWS Availability Zones but on a global scale.

Read More: https://zllovesuki.personal.git.fm/essays/2018/03/andromeda-the-pursuit-of-availability/

Services to be migrated:

• SlugSurvival Web
  • [X] Access (access from Andromeda instead of fmt01 Powered by Cloudflare Workers Site)
• SlugSurvival Data
  • [X] Course Fetcher (save to Andromeda instead of S3)
  • [X] Course Data Access (access from Andromeda instead of S3)
  • [ ] Dispatcher (Messier instead of fmt01)
• SlugSurvival Notify
  • [X] Course Data Access (access from Andromeda instead of S3)
  • [X] Notify Data Access (access from Messier instead of fmt01)
  • [X] User Database (access/save from Andromeda instead of fmt01)
  • [X] Worker (Messier instead of fmt01)
• SlugSurvival Tracking
  • [X] Course Data Access (access from Andromeda instead of S3)
  • [x] Tracking Data Access (access from Messier instead of fmt01)
  • [ ] Dispatcher (Messier instead of fmt01)
  • [ ] Worker (Messier instead of fmt01)
• Documentation
  • [X] Open API (access from Andromeda instead of S3/fmt01)

Other considerations:

• Performance
  • [X] Cache

[zllovesuki]

Current workaround for Web is via Cloudflare. Tracking is currently looking for an usable distributed cron jobs.

[zllovesuki]

CloudFlare released Workers Site, which is perfect for SlugSurvival's JSA.