Show HN: Durable Python Workflows | Hacker News Reader

Show HN: Durable Python Workflows

github.com

64 pointsitayd14 comments

> In addition, it is a durable execution platform for long-running and reliable workflows. It is based on Temporal, hiding many of its infrastructure and coding complexities.

You only hide it for as long as things go smoothly. When a thing breaks then ur hosed. Having seen people try to self host cadence/temporal (not super easy) this makes me skeptical.

In previous life I've been self hosting Cadence and it worked quite well. We then moved to Temporal cloud which was even easier.

That said, if you use autokitteh.cloud, you don't need to worry about Temporal as we're doing all that management behind the scenes.

If someone is already using temporal cloud wouldn’t they be able to use oss AK and just point it at their cloud instance?

yup!

My thoughts as well.. Libraries where the only reason to exist is 'hide this other complex library in the cupboard so you don't have to learn it' almost never achieve that goal, and are generally harmful to net complexity imho.

Hiding is actually not the main thing here. AK allows you to "deploy in a click" instead of deploying the actual workers. We also provide integrations with built in authentication for external services such as Slack, JIRA, etc.

What is so hard about self-hosting Cadence or Temporal?

so last time i used temporal, it has 3-4 separate "services" that need to be deployed in a cluster configuration (not 1 instance) and then they are dependent on a performant cassandra cluster. its certainly not easy.

I didn't want to question it, because I have no experience with it. So thanks for the info.

Hey, author here. We love durable functions and the ease it provides for building stuff. We provide an OSS as well as SaaS "serverless" platform to easily get started with durable functions using vanilla python.

Please take a look as well at https://github.com/autokitteh/kittehub which is full of examples. Also you can use https://autokitteh.cloud for free with some limits.

take a look at https://github.com/conductor-sdk/conductor-python which is easier and will not force you to write with specific framework.

how does it compare to DBOS?

I am sorry i dont understand, what exactly does this library do? Is it a competitor to rapidAPI that lets you quickly integrate APIs in your python app

hi - this is not a library. this is a platform that you upload vanilla python code to it and it makes it run in a durable manner.

Reply on news.ycombinator.com