COVID-19 contact tracing on the BEAM
summary
Built Hygeia on the BEAM using Elixir, Phoenix LiveView and Surface, Ecto with Postgres triggers, Docker/Kubernetes, gen_smtp plus a custom email spool, and Sentry for error collection; first release ships in one month with three developers and users migrate from Trello in under three days. System handles ~400 new index cases/day (~10,000 cases/month), grows to 60,000+ cases and 185+ tracers, and cuts tracing time from ~2 hours to ~15 minutes. Key fixes include custom Postgres versioning, super‑user QA and Kanban workflow.
AI-written summary · about
chapters (37)
- 0:00Intro
- 1:30About our company
- 2:08Agenda
- 3:18Context tracing
- 3:57Why contact tracing
- 5:18Automating contact tracing
- 6:30Team
- 6:54Data
- 8:17Tooling
- 9:50Communication
- 10:29Data aggregation
- 10:47Information security
- 11:04Velocity agility
- 11:23Staying efficient
- 11:54Building the system
- 13:35Code Quality
- 13:59Data Protection
- 14:26Initial Release
- 14:46Adoption
- 15:43Release
- 19:38Technical challenges
- 20:46Liveview encapsulation
- 22:03Liveview parameters
- 22:46Versioning
- 23:36Key takeaway
- 23:53Bad design decision
- 24:54Normalisation
- 26:24Reputation
- 27:06Error Reporting
- 28:13otp supervision
- 29:25jays opinion
- 29:47where are we today
- 30:09the problem really lies with
- 30:34some learnings
- 32:39the future
- 33:11Questions
- 35:36Why kubernetes
related talks
25mGreen Benchmarking of BEAM Languages
32mProcess Data Streams Efficiently in Elixir
38mRapidly building with Gleam
31mEffortless, standardised homelab observability with eBPF
15mProtecting Corporate Accounts with LiveView realtime dashboard
27mBatching Was Yesterday: Real-Time Tracking & Analysis For 100+ Million Visitors
30mAbove the Clouds: Investigating Systems from 32,000 ft.
32mSolid Ground