Thu, Dec 3
Could the existing SVG mechanic be kept as a fallback that only runs for IE11, with clip-path used for other browsers?
Wed, Dec 2
Tue, Dec 1
Thumbor is neither stateful nor high I/O.
Mon, Nov 30
I can't manage to add transfer size to that dashboard, for some reason the time-shifted graphs don't work with it. Maybe a Grafana bug?
I've added cache response type to the per host dashboard: https://grafana-rw.wikimedia.org/d/M7xQ_BeWk/response-time-by-host
The new cacheReponseType field is being collected correctly in Hive:
It's now complete: https://github.com/gi11es/thumbor-docker
Fri, Nov 27
Git tag has been created. I've submitted the new version for Firefox.
Wed, Nov 25
Fixed, I'm currently backfilling the data that went missing during the outage.
Thank you for the detailed update!
This fixed the error, but now I see that the process can't write to graphite:
Tue, Nov 24
The fix appears to work, I see new data coming into Grafana.
I pulled the November data for the experiment, for reference:
FYI, usage of Google Payments for my work account has to go through legal review, which is going to delay my ability to push this update to the Chrome store.
@srodlund it does! I've fixed the article's slug as well, to match the new title. You can go ahead and re-publish it. Thanks!
Mon, Nov 23
@srodlund one thing that might need to get fixed, title included, is that product refers to this feature as "Page previews". That's the name of the product, if you will, which is different than the name of the MediaWiki extension.
They are only shown when $wgGlobalWatchlistDevMode is true, which is true for the beta cluster but false for prod (or will be). Is that enough?
On the latest Chrome stable I see an OPTIONS CORS preflight request before every API request to other domains:
My recollection of why we don't serve user-submitted SVGs directly as thumbnails is that the last time this was looked at there was no robust and up-to-date FLOSS SVG sanitiser that could ensure that the SVGs were safe to display directly in the browser.
Fri, Nov 20
I signed up for Firefox with email@example.com
It's probably going to happen as part of the follow-up to T267327 next quarter, which will be to migrate Thumbor to Kubernetes. I already have it running on Docker with Buster/Thumbor 5/Python 3.
The problem isn't as much the amount of SVGs we get per day, than the fact that we render thumbnails on demand when they're for a file/size combination never requested before. Any extra rendering time is a penalty for that viewer. The issue compounds if they request a lot of new thumbnails at once, making them more likely to run into throttling limits, resulting in erroring images. That can easily happen on galleries that get visited very rarely.
Thu, Nov 19
If the preferences page is extensible (via a hook, maybe?), I think it would make a lot more sense to have these settings added to the "Watchlist" tab there rather than having a dedicated new page.
You'd need to consult a designer from staff on this, this is not my speciality. There is a consistency issue here, we can't have every extension have a custom settings page that behaves completely differently than the wiki's or other extensions.
Not a performance concern, but I notice that the settings page is a navigation dead end. Once you get to it, there is no link back to the Special:GlobalWatchlist page, you're forced to use user history. And once you save changes, there's no link to either Special:GlobalWatchlist or Special:GlobalWatchlistSettings in its default state.
@Krinkle where are the credentials for the chrome-wikimedia-debug Google account and the Firefox equivalent stored? I don't see any mention of it on the repo README or on the wiki page about the extension.
Tue, Nov 17
Mon, Nov 16
Yes, I believe it's a logo that was updated while my patch was in review and that I missed when I updated my patch for the last time before merging (and tried to integrate all logos that changed while my patch was in review).
Mon, Nov 9
Likely caused by T267006
Thu, Nov 5
Progress! Turns out I should have been using pytest instead of nosetest to run the tests. I've filed a small PR to make the python binary used in the Makefile configurable: https://github.com/thumbor/thumbor/pull/1328
FYI, it doesn't:
Seeing that there were a lot of missing or outdated Python Debian packages in Buster compared to what the latest Thumbor master needs, I decided to install the minimum via Debian and all the Python dependencies via pip. Next I will look into whether I can turn all those pip calls into local wheel files, which would be a cleaner way to package this for production.
Awesome, thanks for filing all these tasks. That's all for now, since you mention that the design might change, please ping me when that's the case, so I can look at the changes. If there's any other major change, please move the status of this task back to "Open" and point me to what needs to be looked at.
I was actually already logged into Beta Meta. I've just looked at the error code returned by the API call and it's more explicit: email_not_confirmed.
It might be useful to set up some alerts on that dashboard, to be notified about incidents like this in the feature without having to keep on eye on the dashboard regularly. We can show you how our Grafana-based alerts currently work.
Nov 4 2020
I just took a look at your dashboard, which I haven't consulted in a while. I see that there were some really expensive calls in the last 24 hours:
Changing the status since I haven't heard back about my initial feedback, feel free to make this task "Open" again once you've had a chance to respond to my questions.
Going to the Special:AppManagement page, I see that a progress bar appears immediately and for a very short time:
Nov 3 2020
I ran it manually and it completed successfully.
Unrelated to the performance review, but there seems to be a current limitation in ipv6 address comparison and 0-padding that I saw on the test wiki:
It worked, thanks!
@elukey still the same. Shouldn't it be analytics-deploy?
@elukey the scap deploy fails on permissions, does this look familiar?