Avatar by Buch.
User Details
- User Since
- Nov 14 2014, 10:42 AM (595 w, 3 d)
- Availability
- Available
- IRC Nick
- phuedx
- LDAP User
- Unknown
- MediaWiki User
- Phuedx (WMF) [ Global Accounts ]
Yesterday
Re. the tables in the "From other prior cleanup tasks" table:
Fri, Apr 10
Also in the cons column:
Wed, Apr 8
Another instance of Active Browsing Session Length code causing issues: T422001: '.performer.active_browsing_session_token' should NOT be shorter than 20 characters
I don't wish to speak too soon but I think we've seen the last of both warnings that @dancy mentioned above: https://logstash.wikimedia.org/goto/ee3aee2cf67d6b96eb7274a3899aecb9
Tue, Apr 7
Fri, Mar 27
We're seeing a large number of event validation errors for experiment-related analytics events. The validation error message looks like:
I've filtered out validation errors of this type from the Eventgate validation errors dashboard so that folks don't (1) re-report this issue; and (2) don't think that there's something wrong with their experiment.
Thu, Mar 26
Are we using wiki as a proxy for geo? If so, could/should it be combined with the timezone offset that we captured?
Wed, Mar 25
QuickSurveys supports third party (external) surveys.
We could either create a new test inspired by the existing two tests or (recommended) generalize the test in core so we can share it and improve it for all repos going forward
@Sfaci: You will have to request your change to be deployed in a Puppet request window.
@Sfaci: I like the idea of tidying up repos/data-engineering/metrics-platform separately.
Fri, Mar 20
Of course, I would prefer that WikimediaEvents depends on TestKitchen (see T419679: Make WikimediaEvents depend on TestKitchen) but we can fix this for now.
From the CI logs for the second patch:
Thu, Mar 19
Tue, Mar 17
I met with @hashar about this and this is what we'd have to do to make this work. As I understand it, we'd have to:
Firstly, there are two extensions involved in sending stats: WikimediaEvents and EventLogging. WikimediaEvents is responsible for listening to messages on stats topics, batching them, and queuing beacon requests that send them from the client (the browser) to the server. EventLogging is responsible for the queue.
@cjming: I've responded on Gerrit as well as leaving some other comments to think about. In short, we currently have Experiment#setSchemaID(), which should be enough to get the initial implementation of this deployed. The title of this task may well include "schema" but I'm not sure that we understood what would be involved in implementing that when we estimated this task a 5. Let's discuss the options as a team during standup?
Mon, Mar 16
Mar 12 2026
Mar 11 2026
LGTM!
Experiment Platform are currently running experiments to prove/disprove the hypothesis that the DefaultEventSubmitter and BackgroundQueue mechanisms are sources of event loss. If they aren't, then EventLogging can and should continue to use BackgroundQueue and Test Kitchen can use the equivalent. Sharing the code would complect the two codebases in unnecessary and undesirable ways.
Mar 10 2026
I'm investigating this as part of T409106: X-Experiment-Enrollments EventGate handling reinforcement for MalformedHeaderError cases.
Caused by me disabling the MetricsPlatform extension yesterday. Being bold.
@JVanderhoop-WMF: I've been bold and prioritised this as High. We appear to have missed disabling this warning somewhere and should disable it quickly.
@JVanderhoop-WMF: I've been bold and assigned this to you. We'll need copy for the option. I'm not sure how you want to approach writing it.
Mar 9 2026
Quoting myself from Slack:
I'm seeing flowing on the eqiad.product_metrics.web_base stream.
