Selwin Ong
Selwin Ong
Sorry for the late reply, I'd love to enable or build more monitoring hooks into RQ. Could you be more specific on how we should proceed?
Sorry for the late reply. Thanks for bringing this up. The non atomic movement from queue to registry is also an issue that I wish to solve because a job...
> One possible mitigation would be to have the main worker process mark the job as started immediately after dequeuing, rather than waiting for the horse to start. This would...
Just a quick update. I'm experimenting with using `LMOVE` to prevent jobs from being orphaned. This is the clean way of implementing reliable queue, but it only works if the...
> Our workers are working on several queues — but I'm not sure to understand the limitation of LMOVE in this case? Redis' [BLMOVE](https://redis.io/commands/blmove/) command does not support specifying multiple...
That's because when you run `rq worker foo-queue bar-queue`, it will try to listen to incoming jobs from two queues. Internally, it uses [this command](https://github.com/rq/rq/blob/89fa8ae0b560ccdaeb54d173dfbea6577a4b2163/rq/queue.py#L1291): ```python redis.blpop('foo-queue', 'bar-queue') ``` When...
The intermediate queue is introduced to prevent a job from getting "orphaned", which is a condition where the job hash is present in Redis, but the job ID is not...
> Now, you introduced the intermediary queue only for single-queue workers, which I understand (you explained in details above why), but I think it could also work with multiple queues...
Hey, so it seems like [_set_connection(connection: Redis)](https://github.com/rq/rq/blob/master/rq/worker.py#L411) does not work with RedisCluster connection. Unfortunately I don't use Redis cluster, nor do I have access to a dev environment with Redis...