What is a Task Queue?

Rather than executing a task the instant it's requested, a task queue decouples task creation from task execution: a producer enqueues work (a job to run, a message to process), and separate worker processes dequeue and execute tasks independently, often with concurrency control (how many run at once), retry logic for failures, and priority ordering. This pattern is essential for handling load spikes gracefully — incoming requests get queued rather than overwhelming the processing layer — and for long-running work that shouldn't block a synchronous request/response cycle. Reliable task queues need to handle failure modes explicitly: what happens if a worker crashes mid-task (does the task get requeued, or lost?), how many retries are allowed before a task is marked permanently failed, and how to prevent the same task being processed twice by two workers racing on the same queue entry. These guarantees (at-least-once vs. exactly-once processing, idempotent task handlers) are what separate a production-grade task queue from a naive in-memory list.