Lessons from building a background task processing library
This talk traces the evolution of a Rust background task processing library, through multiple versions spanning five years of development. The presenter examines key architectural decisions and their rationale, including the implementation of workers as actors and the dual approach to task definitions using both trait-based and trait-free APIs. The presentation explores how dependency injection creates an ergonomic developer experience, the integration of middleware for extensible task handling, and solutions for production concerns like graceful shutdown and rolling restarts. Attendees will gain insight into building complex background task workflows and understand the trade-offs behind API design choices. By analyzing version-by-version progress, the talk provides practical lessons in library design, backward compatibility considerations, and evolving APIs based on community feedback. Developers working on or considering background job processing in Rust will learn from real-world architectural decisions in a production-ready system.
