Idioms for building fault-tolerant applications with Elixir
summary
Elixir builds on the Erlang VM and demonstrates concrete idioms: lightweight processes isolated by message passing, supervisors that restart failed workers, and supervision trees that give live introspection via the observer / LiveDashboard. Livebook and Phoenix demos reproduce WhatsApp’s claim of ~2 million connections on one machine after removing message-queue bottlenecks, and Pinterest reports ~95% fewer servers and ~$2M/year saved. Practical tools shown include Ecto for queries, assertions as macros in tests, Nx for GPU numeric work, and Nerves for supervised embedded systems.
AI-written summary · about
chapters (10)
- 0:00Start
- 5:22Functional programming
- 14:53Idioms
- 15:08Sequential code
- 28:11Visibility of the application state
- 30:51Compatibility
- 31:14Extensibility
- 31:37test "basic operations" do
- 35:33Productivity
- 35:50First-class documentation • Tooling (ExUnit, lex, Mix) • Hex packages
related talks
44mIdioms for Building Fault-tolerant Applications with Elixir
53mIdioms for Building Distributed Fault-tolerant Applications with Elixir
03h 19mWorkshop - All Aboard The Elixir Express!
44mElixir & Phoenix - Fast, Concurrent and Explicit
01h 06mWhat every Node.js developer needs to know about Elixir
53mElixir: a potion to mix complexity and concurrency in real‑time
19mElixir – what it is and why is it amazing?
41mElixir & Phoenix – fast, concurrent and explicit