Hello, is there any plan on making the orion ui si...
# prefect-ui
r
Hello, is there any plan on making the orion ui similar to prefect 1.0 UI on the cloud?
k
Depends what you mean by similar, do you mean us hosting it?
r
yep and the visual aesthetics etc. i liked the layout on the cloud better
k
That had too much overhead and was not performant so we went with a more lightweight thing, Orion is way more flexible though and has more features. The hosting, yes we’ll have some form of a hosted version soon
r
oh i see. sounds good then.
z
The Orion UI is definitely incomplete right now so don’t pass judgement too early!
We’re dedicating a lot of time to improving the UI experience.
r
is that anyway one can view logs of errors on the UI? i see the task running but on the backend i've gotten a 1006 error due to some tcp connectivity. However, the prefect orion UI still shows the flows constantly running and im not able to see that an issue has occured
z
We’ve added extensive “crash” detection that will be out with the next release. If a task/flow fails outside of your code, we’ll report the traceback in the logs and update the state accordingly.
r
awesome
when is orion going to be cloud ready?
makes it easier for my team and for us to share pipelines among us
z
In some cases, your flow/task can crash and be unable to contact the API and notify about the failure state. To resolve this, we’ll add some form of healthchecks/heartbeats but these have been a huge pain in V1 so we want to design them carefully and haven’t added them yet.
r
how about an if an error occurs send email option?
or would that be done on the backend on prefect through ur package
z
That sort of crash is usually an aggressive shutdown by the OS or a complete loss of connectivity, so we need a way to detect that it is not running without it phoning out.
I can’t share a date for cloud 🙂 it’s on the way though
r
wouldn't 1006 count as an aggressive shutdown by the os?
z
Where did it occur? A connection closing doesn’t necessarily mean the whole flow is going to get shutdown.
By aggressive shutdown I mean that the process exits without going through Python’s exit handling mechanisms.
r
well im streaming data via websockets in python. On the backend i wrote a try except clause to throw out any errors locally before i deployed it on the aws cloud. I checked the backend logs after i stepped out of my computer for couple hours and noticed lots of messages
i keep getting this error
Copy code
ot:no close frame received or sent
INFO:root:exception occured at ftx-L1
INFO:root:Socket Closed: at 2022-03-10 17:02:51.352759;
            close reason: ,
                close code: 1006