Node.js Queue Scheduling: Webhook Retries, Delayed Jobs, Cron, and DLQ Recovery

The article argues that a healthtech SaaS worker pool should rely on a durable queue with explicit leases, bounded retries, a dead-letter queue, and a drain mode, with cron serving only as a trigger. It uses a scenario of appointment reminders sent through a public webhook, where a downstream provider's rate limit drops during an incident. The author stresses that queues counting only successful requests can hide growing backlogs, so tracking job age separately is essential. The recommended runbook has operators pause intake, drain workers one pool at a time, and avoid tight retry loops by honoring Retry-After headers and adding jitter. The piece concludes that incident recovery should be procedural and measured, with attempt events exported and the incident closed only after a clean, deduplicated cron run.