ReadMeFeed / Community knowledge
Real questions. Useful conversations. Find the people who know your stack.
Developer tools · #erlang
Discover how Erlang’s OTP supervision trees and the "Let It Crash" philosophy let you build self‑healing distributed systems without defensive error handling.
Learn how to build a minimal OTP supervision tree in Erlang: requirements, design, trust boundaries, operational checks, failure modes, and when to redesign.
An architecture note on wrapping Erlang libraries from Gleam: private @external declarations, validating wrappers, tagged-tuple error contracts, and when to switch to OTP actors instead.
How OTP supervisors restart failed workers: one_for_one vs one_for_all vs rest_for_one, a worked supervisor module with map child specs, intensity/period escalation, and the state-loss and crash-loop mistakes to avoid.
A concise architecture note that outlines requirements, the minimal supervision‑tree design, trust boundaries, operational checks, failure modes, and triggers for redesign when using Erlang/OTP for resilient services.
Supervision strategy is a claim about which Erlang processes depend on which. Pick it from the dependency graph, then confirm it with a deliberate kill on a test node.
Decide whether to compile Gleam to the Erlang BEAM VM or to JavaScript by weighing concurrency, fault tolerance, and deployment constraints. A side‑by‑side table, trade‑off analysis, and a concrete HTTP endpoint test guide help you pick the right target for your service.
Erlang’s lightweight processes enable millions of concurrent tasks with minimal memory. Learn how to build a fault‑tolerant counter server, watch it recover from crashes, and understand key trade‑offs like mailbox size, data copying, and supervision loops.
A technical decision guide for Erlang/OTP state management, comparing ETS, persistent_term, and the process dictionary based on latency, concurrency, and mutability.
gleam