Kue: Future of kue

Created on 15 Aug 2017  Â·  10Comments  Â·  Source: Automattic/kue

Hi everybody,

I use kue in various projects for French government.
I open this issue because I need a bit of visibility on the maintenance and the next evolutions.

Some points:

  • @behrad is announcing a v1 since more than one year. As I can see on the v1 branch, it is still a WIP. It blocks any ambitious PR. How can we help to target a quick release?
  • Parts of the library are staled (UI, API).
  • Other similar projects (as https://github.com/agenda/agenda) are pretty active.

I have the feeling that the actual kue is too big to be efficiently embedded in high-level libraries, and a bit too small to cover advanced use cases.

In the past I slightly modified kue and published kue-lite for our internal usage, and we are now working on a wrapper around kue which support job unicity, promises, add advanced logging, job configuration with YAML or JSON file and out-of-the-box jobs worker. Maybe some of these features would have their place in kue.

It could be interesting to have a package kue-core with just the logic and connection to Redis, and a full-featured kue with JSON API and UI. And why not kue-ui, kue-websocket, kue-api, kue-webhooks…

Feel free to discuss my position ;)

Most helpful comment

I was very busy for months... I will try to do my best to clean up pending PRs...

All 10 comments

Don't! Agenda is not production ready.

When considering what to use I found agenda and kue. After reading a little bit the issues agenda has, I wouldn't recommend it for production right now. It has a severe memory leak issue

That's the main poin's the main reason I'm giving kue a try.

EDIT: this is a response to this @niftylettuce comment (sadly, he removes his comments just for fun):

@gentunian I've been using Agenda in production for many years doing hundred of thousands of transactions. It's definitely 100% production ready. Did you even visit the repo?

@gentunian you are making inaccurate assumptions without knowledge and deep understanding and insight.

to others, please take any comments here with a grain of salt.

I've read all what I needed yo read. The api, docs, and the huge open issues list.

It's enough for me to not use It. Kue has a leaky feature for searching too and has been solved by disabling default behaviour. So, until the memory leak is not solved I won't use agenda un production.

Please, don't make assumptions yourself about what I know and what I read, my salty boy.

This is a response to @niftylettuce post (sadly, he removes his comments just for fun):

@gentunian last time I checked you have zero GitHub contributions. Thanks for the kind words.

So? What assumptions will you make now about that?

And finally, I will leave this for reference: https://github.com/agenda/agenda/issues?q=is%3Aopen+is%3Aissue+label%3Abug

I won't step into agenda until critical bugs are fixed. I consider some of those critical.

My point is not comparing Kue with Agenda, but Kue should be as many active as Agenda.

I was very busy for months... I will try to do my best to clean up pending PRs...

Maybe I can help?

@jdesboeufs Sorry to resurrect this long-dead issue but I'd be very interested in using your improvied version of kue. Any chance you will open-source it?

Ça serait super ;)

Hi @vroudge

We finally switched to bull.
https://github.com/OptimalBits/bull

I stopped using Bull, and just released Bree today.

Docs/Site: https://jobscheduler.net
GitHub: https://github.com/breejs/bree

Was this page helpful?
0 / 5 - 0 ratings

Related issues

syadykin picture syadykin  Â·  8Comments

mogadanez picture mogadanez  Â·  3Comments

narainsagar picture narainsagar  Â·  8Comments

JperF picture JperF  Â·  8Comments

danpalmer picture danpalmer  Â·  8Comments