atlas-queue

High-performance distributed task queue built in Rust.

Atlas Queue is a production-inspired learning project for building queueing, worker, scheduling, persistence, and distributed systems fundamentals in Rust.

Current Phase

Phase 4: Postgres Persistence & Crash Recovery

• Persist tasks in Postgres.
• Dequeue safely across concurrent workers with FOR UPDATE SKIP LOCKED.
• Reclaim running tasks whose leases expire after worker failure.
• Track task lifecycle through typestate at the Rust API boundary.

Run

docker compose up -d
set DATABASE_URL=postgres://atlas:atlas@localhost:5432/atlas_queue
cargo run

Local Postgres connection string: postgres://atlas:atlas@localhost:5432/atlas_queue

Test

cargo test