[ field notes ]
What we've learned running Laravel queues in production — concurrency, Horizon, Redis, and the operational edges that only surface under real traffic.
· 16 min read
"60 requests a minute" and "no parallel requests" are two different limits, and Laravel solves them in two different places. A tour of every option — RateLimited, WithoutOverlapping, funnels, single-slot supervisors — the release churn they all quietly cost you, and how to let urgent jobs jump the line without breaking the limiter.
Read more· 11 min read
Most job bugs aren't logic errors — they're assumptions about when and how many times a job runs that hold in development and fall apart under load. Twelve practices that keep a Laravel queue healthy in production, ending with the deploy-time rule that no linter catches.
Read more· 10 min read
Every Horizon supervisor forces a choice that rarely gets discussed: turn balancing on and pay for a floor of idle workers, or turn it off and let a busy queue starve the ones beneath it. Skyline's weighted queues collapse the two bad options into one good one.
Read more· 12 min read
Mark a job ShouldBeUnique and move on — that's the theory. In production, ShouldBeUnique and WithoutOverlapping solve different problems and fail in different ways: dropped jobs, exhausted attempts, orphaned locks, and invisible bottlenecks.
Read more