#169 [New Project] Static Dashboards
Opened 8 years ago by tflink. Modified 6 years ago

While resultsdb is very useful, it would be very useful to have pretty, easily discoverable "dashboards" which show computed results from a number of individual results.

The first use case which comes to mind is moving the OpenQA triggering and results into taskotron. We'd want to have top level results like "KDE Live ISO" or "Workstation ISO" computed from multiple cases like base_selinux, basic_service_manipulation and default_install.

To hopefully keep this as simple as we can, have a form of input files (yaml, probably) which list a top level "name" which has a list of testcase names underneath it. When there's a new milestone that we're interested (daily compose, TC, RC, alpha etc.), the static generator is run, creating a new set of static HTML pages which represent a snapshot of the current state of tests.

One possible extension would be to have json files in addition to the html. The idea here would be to have a fake restful service - another program could do a GET on something like <html host>/<snapshot identifier (compose, timestamp etc.)>/<top level thing>.json and we could all pretend that it's not just a bunch of static HTML.

My first thought for a design is depicted below. The big points are:
* breadcrumbs showing where you are in the tree (shapshot identifier, group, top level thing etc.)
* an "easy to figure out at a glance" colored list of green and red (maybe yellow) results
* a concept of test cases which are non-blocking - if they fail, they don't cause the overal higher level thing to be a failure as well

{F45673}


! In T725#12401, @tflink wrote:
I've gotten a very basic start on this:

https://bitbucket.org/tflink/taskdash

It's pretty much some examples ATM but figured I would mention it

Metadata Update from @kparal:
- Issue tagged with: task-idea

6 years ago

Metadata Update from @kparal:
- Issue untagged with: task-idea

6 years ago

Login to comment on this ticket.

Metadata