Hi, I'd like to implement retry with an exponential backoff as used in Sidekiq, that is if a job fails it is requeued with an ever longer pause.
Right now if I NACK a job, it returns to the queue immediately. My other option is to ACK the job and add it again, with a delay. This works OK, but from client's point of view it changes the job ID and that makes tracking failed jobs more difficult.
The NACK command with a delay support could look like this:
NACK <job-id> ... <job-id> [DELAY <sec>]
What do you think?
Hi, I'd like to implement retry with an exponential backoff as used in Sidekiq, that is if a job fails it is requeued with an ever longer pause.
Right now if I NACK a job, it returns to the queue immediately. My other option is to ACK the job and add it again, with a delay. This works OK, but from client's point of view it changes the job ID and that makes tracking failed jobs more difficult.
The NACK command with a delay support could look like this:
NACK <job-id> ... <job-id> [DELAY <sec>]What do you think?