Functions time out, cron runs once a minute at best. For jobs like "resize 200 images after upload", what are people reaching for? Queues, workflows, a separate worker somewhere else?
Functions time out, cron runs once a minute at best. For jobs like "resize 200 images after upload", what are people reaching for? Queues, workflows, a separate worker somewhere else?
Queue plus a function consumer. Each image is one message, so nothing gets near the timeout.
We moved long jobs to the durable workflow beta — retries and steps are built in, which saved us writing our own state machine.
Still a cron that drains a table here. Boring, works, cheap.