Re: KernelCI Dashboard - Hardware Listing Behavior
Ben Copeland <[email protected]> Thu, 22 Jan 2026 10:45:42 +0000
| Newsgroups | dev.linux.lists.kernelci |
|---|---|
| Message-ID | <CAL0q8a5s0krCxSNjiCxQ=4piWFYX0at1iMCtCBwiwxQZ0NKXdQ@mail.gmail.com> |
On Wed, 21 Jan 2026 at 20:54, Gustavo Flores <[email protected]> wrote: > > Hi everyone, > > We are working on some improvements on the dashboard and would like to > hear your opinion. > > Currently, the hardware listing and details pages on the dashboard > only show hardware that has tests in the latest checkout of each tree. > As a result, some hardware ends up being effectively "hidden" from the > views. How much hardware is being hidden? And do we know why? > > We discussed this internally and see two possible approaches: > > 1. Show all tests within the selected time frame (e.g. last 5/15/30 > days), regardless of checkout. > 2. Show tests from the most recent checkout in which each specific > hardware was tested, even if it's not the latest checkout of a tree. 2. to me makes sense, but it might confuse the user if they click down and don't see the checkout that caused the failure. I am wondering why some hardware is not running on some boards for so long that it goes missing. Is this down to board issues, or is it by design that they are running these checkouts? Seems odd behaviour nevertheless! Ben > > With the first approach, the number of displayed tests would likely > increase for larger time windows, since we would no longer be > filtering by checkout, but this shouldn't be a problem. > > Since this would change the current behavior of these pages, we=E2=80=99d= like > your feedback: which approach do you think better fits your needs? > > Best, > Gustavo Flores >