:thread: What are your primary goals or tasks when...
# prefect-ui
b
🧵 What are your primary goals or tasks when using Prefect 2.0? How can the dashboard help you achieve them more efficiently?
t
Most of our flows start multiple sub-flows, which clutters the current dashboard quickly and makes it hard to find the parent-flows. Maybe sub-flows could be collapsible into the parent-flow?
âž• 3
đź‘Ť 1
s
On the subject of subflows - I think the UI could have a better approach to linking subflows to their parent flows - e.g., subflows show up as separate flows in the UI flow runs view w/o any context/metadata showing that they are a subflow
đź’Ż 5
v
I'm not sure which thread to put this into, but I would really like the ability to cancel flow runs directly from the flow run view!
đź‘Ť 1
r
I would like the UI to remember my tag-filter so I don’t have to select it every time
âž• 2
My main goal is “See flow runs that are relevant to me (especially failed ones), then see their logs”. Currently this works ok but the filtering part could be better - including remembering filters (as mentioned above) and being able to filter by state. edit: looks like you can already filter by state, for some reason I missed that facepalm
Also if there is non-log diagnostic info (eg kubernetes events), having quick access to that would also be helpful
Also the way the flow runs move around as you hover over them (to show the checkbox) feels glitchy and annoying. In generally the dashboard UI feels a bit “sticky” or glitchy
Also all this is coming from a place where I haven’t spent a long time using the 2.0 UI yet (we’re still mid-migration from 1.0)
b
Thank you @Ryan Sattler!
k
Hi @Brandon Reid, hope you are well. I use tasks to manage data pipeline for ETL and model training/ updates. For me the dashboard is the starting point every morning to access what didn't work and where I should focus if something requited fixing
đź‘Ť 1
b
@Kirill Egorov in order to do that, are you primarily looking for failed flow runs?
k
So it is failed flow runs -> failed tasks -> logs. Quicker the process to find, identify, retry or fix is better for me
🙏 1