GigCityElixir23 - Bryan Hunter
Why Hacker News recommends it
HN recommends it as a rare real-world look at a large fault-tolerant Erlang/Elixir system with no downtime in healthcare.
What it is
Bryan Hunter's GigCityElixir 2023 talk on HCA Healthcare's Waterpark system, built on Elixir/Erlang. It describes how the architecture is distributed and fault-tolerant, with a multi-datacenter cluster and runtime code swapping.
Recommended by 4 HN commenters across 5 different months, 2023 to 2025.
What Hacker News said
Strong rec rramadass · 2023-12 · on “Ask HN: Good examples of fault-tolerant Erlang code?”
Not code but an excellent presentation of design/architecture of a Real-World fault-tolerant and distributed System in Erlang/Elixir - [this video]
Strong rec rramadass · 2024-01 · on “The Erlang Ecosystem (2018) [video]”
After watching the above which gives one an idea of the Erlang System's capabilities, watch this presentation by Bryan Hunter where the design and architecture of a complete distributed and fault-tolerant Erlang/Elixir system is shown; You get to really see the mind-blowing power of the Erlang System in this - [this video]
Recommends notpublic · 2023-06 · on “A LiveView Is a Process”
This video may be of interest to you: [this video] Bryan Hunter explains how they do it at HCA healthcare (not specifically LiveView process, but processes in general).
Recommends GCUMstlyHarmls · 2024-12 · on “Elixir/Erlang Hot Swapping Code (2016)”
This is a talk about a large scale, resilient elixir/erlang deployment in healthcare. Specifically they talk about running with no down time using hot code reloading here: [this video] but the whole talk is quite interesting regarding availability. Warning: the video is quite quiet.
Recommends chantepierre · 2025-07 · on “Leveraging Elixir's hot code loading capabilities to modularize a monolithic app”
To add to this topic, people who do not know about erlang's hot code loading should watch this talk : [this video] A multi-DC running cluster where parts are progressively swapped at runtime. No database, only OTP.
Cites darkmarmot · 2023-08 · on “Unpacking Elixir: Concurrency”
We use it across multiple versions of our software running in the same cluster. As long as you dark launch API changes, it's not much of an issue. [this video] (Doh, just realized I was replying to Mr. Elixir himself! And you're familiar with our project anyway :)
Cites darkmarmot · 2024-08 · on “Constraining writers in distributed systems”
We use something like this at scale, similar to Riak's design. We use it to store things like the active state of millions of hospital patients in RAM for high availability (zero downtime when DCs or nodes fail). One copy per data center with a writer at one of the DCs. Our current cluster has 32 nodes (8 per DC -- we should have 5 DCs across the US but only have 4 at the moment). You can learn more about it here: [this video] .