Panel: Elixir Adoption
summary
Legacy PHP stack migrates to Elixir with a concrete proof: an autosuggest prototype built in Phoenix that routes real traffic and is ramped traffic to 100% after automated tests showed identical results and much faster responses. Teams hit deployment and runtime config pain points like poolboy environment variables and tradeoffs between Kubernetes vs distributed Erlang for broadcasts. Adoption uses lunch-and-learns, incremental endpoints, measurable performance and an anecdote of a forgotten production app idling at 1% CPU.
AI-written summary · about
chapters (17)
- 0:00Elixir adoption panel intro
- 0:28Team buy-in for Elixir
- 1:34Choosing Elixir for rewrites
- 2:16First Elixir prototypes
- 3:20Gradual Elixir rollout
- 4:46Pitching Elixir to teams
- 6:14Lunch and learns for Elixir
- 7:37Convincing with performance graphs
- 8:21Elixir adoption pain points
- 9:46Kubernetes deployment challenges
- 10:28Umbrella apps transition
- 11:21Erlang and Kubernetes integration
- 12:47Elixir vs Java comparison
- 13:55Immutability and concurrency benefits
- 16:25When Elixir is not the right choice
- 17:32Learning curve and juniors
- 18:16Final thoughts and Q&A
related talks
27mHow Elixir is driving Australian product innovation
28mHow Elixir Helped Me to Love the Back End
39mMoving to Elixir from Ruby
36mLessons learned from Elixir Learning Paths
40mBeyond your daily coding
42mErlang from the Outside- 36mLessons learned from Elixir Learning Paths
06mLT: The Elixir Hiring Paradox