Hi @snowman, unfortunately, we haven’t had much time to work on the Python client in the past few months and still haven’t put anything out there (it’s still merely a side project / would be a community contribution for https://awesome.zeebe.io).
Regarding polling and clients in general: we’ve heard feedback that users would prefer a different approach than what Zeebe does currently for a number of (perfectly valid) reasons, and as a first step, we’re looking into a long polling feature. We wouldn’t start on this until at least Q3, but it is definitely on our radar.
Curious to know, are you running into issues with latency and/or resource utilization due to the current polling approach?
Thank you @wints !
Focus on core functionality make perfect sense to me.
Since Zeebe are in developer preview stage; releasing a “unfinished” client might not be a bad idea to draw participation from wider community.
Not running into any issue right now for exploratory use cases.
There is always an element of premature optimizations on my part.
On the other hand; reduce stack complexity is always good IMHO:
There are mention of Python Celery in one of the thread. If I can replace [Celery/RQ/Async Worker] with Zeebe while adding functionality is very attractive. It goes well with the the “battery included” approached.
thank you @jwulf
I was able to follow along the blog/tutorial and get a simple python worker going.
Maybe I am becoming more paranoid in my old age as the whole endeavor was “too easy”
“I don’t know what I don’t know” about what might be missing from a “full feature” client is the reason why I was looking for an official client.