Previously, when changing the measure on evo.html, we would update the selected
aggregates first before pulling down the new evolution data. This would ensure
switching from, say, linear to flag, would mean we would stop displaying the
25th percentile and start displaying false, true, and invalid value percentages.
This was accomplished through calling Telemetry.getHistogramInfo for the
measure on every displayed channel and version, with no filters.
This hit the aggregates.tmo endpoint, asking for the full submission_date
aggregate data for the entire display... which was then ignored except for the
metric's `kind` and `buckets`.
To add insult to injury, this was unlikely to be cached, since we asked for it
filter-free. Usually there is at least an `application=Firefox` filter in place
which means we ask for all of this data, then never reuse it.
This change skips this entirely and pulls the kind and buckets out of the
result we get from asking for the display data.
This shaves anywhere from two to twenty-two seconds off of the time it takes to
switch metrics on evo.html.
Publishing the usecounters dashboard from a separate dashboard broke
the “serve /index.html for /” behaviour.
Fixing the link here until i figure that out.
I assume that Mission Control is currently not as important as a link
target as HW report, Redash & Experiments Viewer.
Moving its link down accordingly.
We don't know where users come from or go to on TMO, and that's an odd sort of
lapse for a data-based web application.
This is preliminary: just gtags for now.
* Move Hardware Report to main dashboards
This dashboard is hard to find, but answers some important questions
for engineers and the broader user base.
* Make hardware report description more snappy
* Highlight & expand documentation links
* Fix column height to avoid scroll bars
* Add "getting help" section
* Minor rewording
* Make clear which items are LDAP protected
* Add experiments viewer dashboard
* Move reports.tmo due to column overflow
* Make the experiments viewer description more snappy
* Change "what is new" to "what's new"
* Change the page title to "Telemetry portal
* Add Shield docs
* Make columns non-scrollable
Make columns non-scrollable, instead make them grow with the page and
let people scroll the page.
* Remove non-existing dashboards
Remove non-existing dashboards, instead add an item that tells you to
file a bug for anything missing.
* Break up lists into separate dashboard-list elements
Now that we don’t use scrollable list anymore, we can just use separate
lists for the different use-cases.
* Make clear that sql.tmo shows dashboards too
* Add Amplitude
* Use "tools" instead of "dashboards"
Not all things we link to are dashboards, lets use something more
generic.
* Remove wiki link
* Remove duplicated bug report link
This is probably triplicated.
* Use more consistent capitalisation
* Remove unused .scrollable-dashboard-list styling
* Remove top border on first dashboard-list items
The format changed from {"hashed_id": "name"} to
{"hashed_id": {"name": "Some Name", "id": "addon guid", ...}.
This commit also makes the recommender only recommend
non-legacy addons.
See https://bugzilla.mozilla.org/show_bug.cgi?id=1344003 for
more context. In short, I created a profiler-like UI for hang
data available on my GitHub Pages page. This change links the
Background Hang Reporting link on the main view to this page.
Chris H-C
o0Ignition0o
Georg Fritzsche
Shikhar Agnihotri
Soumya Himanish Mohapatra
Vanessa Sabino
Robert Strong
Frank Bertsch
William Lachance
Chris H-C
Armen Zambrano G
Mark Reid
Adrian Wielgosik
ab0092
Aditya Addepalli
Alessio Placitelli
Panagiotis Astithas
Doug Thayer