An engine that enables truly global concurrency, global rate limiting, queue partitioning, async execution tracking, true snooze counting, unique bulk inserts, and auto insert batching.
Extend the base worker with advanced options like encryption, enforced structure, and recorded output.
Insert jobs and await the results across any number of nodes, i.e. persistent distributed tasks.
Advanced helpers for testing supervised Oban instances, workers, and making assertions about enqueued jobs.
Process related jobs while tracking overall progress across all nodes and executing optional callbacks.
Process jobs in strict sequential order regardless of scheduling, retries, and queue concurrency.
Process jobs atomically in groups based on size or timeout, with robust error handling semantics and optional partitioning.
Compose workers together with arbitrary dependencies between jobs, allowing sequential, fan-out and fan-in execution.
Configure cron scheduling before boot or during runtime, globally, across an entire cluster with scheduling guarantees and timezone overrides. An ideal solution for applications that must dynamically start and manage scheduled tasks at runtime.
Rescue orphaned jobs reliably, without the possibility of duplicate execution and the ability to retry jobs that have exhausted their attempts.
Optimize query performance, minimize database bloat, and efficiently prune historic jobs by storing jobs in a managed partitioned table.
Automatically adjust job priorities to prevent queue starvation and ensure all jobs are eventually processed.
Delete historic jobs on a flexible CRON schedule using configurable limits with per-queue, per-worker, and per-state overrides.
Configure queues at runtime, persist changes across restarts, and optionally limit queues to certain nodes—ideal for applications that dynamically start, stop, or modify queues at runtime.
Horizontally auto-scale worker nodes up and down to optimize queue throughput during high traffic events, and pare back to zero or more nodes during a lull.