zlacker

[return to "Launch HN: Hatchet (YC W24) – Open-source task queue, now with a cloud version"]
1. krick+RD1[view] [source] 2024-06-28 02:55:53
>>abelan+(OP)
Can somebody explain why would I use it instead a simple Redis/SQS/Postgres queue implemented in 50 LOC (+ some grafana panel for monitoring) (which is pretty much mandatory even for a wrapper of this or any other service)? I'm not trying to mock it, it's a serious question. What is implied by "task queue" that makes it worth bothering to use a dedicated service?
◧◩
2. vasco+DS1[view] [source] 2024-06-28 06:18:21
>>krick+RD1
You can use celery with postgres without issues if you want the stuff you don't get with that, like tweakable retries, tweakable amounts of prefetch and other important-at-scale things. Plus out of the box working sdk with higher level patterns for you developers. Like what if devs want to track how long something waited in the queue or a metric about retries etc, things that you'd have to roll by hand.
◧◩◪
3. mathnm+u12[view] [source] 2024-06-28 08:30:48
>>vasco+DS1
> You can use celery with postgres without issues

How? This issue still seems to be open after 6 years: https://github.com/celery/celery/issues/5149

[go to top]