Skip to content

Conversation

@consideRatio
Copy link
Member

@consideRatio consideRatio commented Nov 11, 2025

Before in JupyterHub dashboard:
image

Before in Cluster dashboard:
image

After in either:
image

@consideRatio consideRatio force-pushed the pr/style-non-running-pods branch from 1c85e1a to 7d2075c Compare November 11, 2025 10:30
@consideRatio consideRatio changed the title Make non-running pods panels consistent and style them intuitively~ Make non-running pods panels consistent and style them intuitively Nov 11, 2025
Copy link
Collaborator

@jnywong jnywong left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Hmm with this one I usually encounter a problem where the colour for another series "overlays" the time series for another, e.g. in the screenshot, the orange line for unknown should be zero over all time, but "overlays" the pending and failed series.

I see that you went with the time series and not bar chart panel to visualise. I wonder if we can overcome the "overlaying" by using a stacked barchart with mode normal?

Image

@consideRatio
Copy link
Member Author

consideRatio commented Nov 13, 2025

Hmmm this is very confusing - why does it show up as orange peaks when not stacking? Shouldnt the orange stay at zero!?

Separately, in your screenshot, i see failed as green - but this PR meant to make it red. Is thr screenshot from using this PRs panel? If so, there is a bug in it as i meant to change those colors!


Edit: it looks like you hover a peak of 4 on the yaxis - 2 failed + 2 pending, but showing up as orange unknown. This indicates a stacked time series - so, the old version of this panel right?

@jnywong
Copy link
Collaborator

jnywong commented Nov 14, 2025

Ah man I had the wrong PR checked out 😅

The same thing happens again however (see screenshot)

Screenshot 2025-11-14 at 09 49 17

@consideRatio
Copy link
Member Author

consideRatio commented Nov 14, 2025

Ah hmm, it seems to me like a visual issue no matter what if we use stacking or not - at least when the situation is that you are looking at data very zoomed out so that the data points are just spikes.

Without stacking With stacking
Screenshot 2025-11-14 at 09 49 17 Image

Is there a case to make it stacked or non-stacked? I perceived that I could better read relevant details by using non-stacked for this panel.

Here is a comparison:

Without stacking With stacking
image image

When using stacking, we "lift" some time series on top of others, so in the example above, it looks like the orange color (unknown) is actually non-zero.

@jnywong
Copy link
Collaborator

jnywong commented Nov 14, 2025

Yeah let's keep it non-stacked in this case. So I think this PR is good to go! Thanks for making these small UX enhancements @consideRatio

@consideRatio
Copy link
Member Author

Thank you @jnywong!!

@consideRatio consideRatio merged commit b212d48 into jupyterhub:main Nov 14, 2025
4 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants