Most teams still default to "add Redis" the moment their app grows a job queue. Wheels 4.0 takes a different position: the queue lives in the database you already have. <PR #1934 — Background jobs> What you get without standing up Redis: •

wheels jobs work

— persistent CLI daemon, optimistic-lock claim so workers can't double-process •

wheels jobs status / retry / purge / monitor

— full operator surface, live TUI dashboard • Configurable exponential backoff per job (

Min(baseDelay * 2^attempt, maxDelay)

) • Timeout recovery — a crashed worker doesn't orphan its job •

wheels_jobs

table auto-creates on first enqueue — no migration step • Tenant-aware enqueueing for free (no

tenant_id

in the payload) Where Redis still wins (and the post says so plainly): hundreds of thousands of jobs/min, pub/sub fan-out, sub-50ms pickup latency. Most apps are nowhere near these limits. <Read the full post the full post>