TaskIQ in production

[tyleralbee]

What does a good production setup look like with TaskIQ?

I did see in another thread that @s3rius suggested using Helm to manage Kubernetes applications, saying this:

Personally I use helm charts for my applications and both taskiq workers and schedulers are declared using the same deployment template, but with different command.

I realize there is no one-size-fits-all answer to this question, but I'd appreciate some advice from more experienced developers. I'm trying to figure out how to host the broker/workers (ideally co-located with Redis) with the ability to scale up/down workers depending on load. In my case, there's one task that can take 5+minutes.

I have enjoyed working with TaskIQ and I appreciate the quality of work put in by the maintainers.

[devraj]

Seeing I am the guy who started that thread, I thought I'd chime in 😀 and add my two cents worth. I have got it all working via a helm chart as described on the thread, it's a normal deployment which you should be able to scale up and down with demand (note that my chart is spinning up a worker, not a scheduler - which shouldn't be difficult but be weary that you shouldn't have multiple of them).

The chart is available here I am yet to document or finalise some loose ends.

If it is of assistance then I am willing to write out my finding in a document and publish it on our experimental repo.

[tyleralbee]

That would be awesome! Your link to the chart is already very helpful. Thanks a ton.