New data should be coming in fine, now, so we're ready for this any time. Thanks much!!!

Hi! We were able to deploy this fine! :) (Thanks much to @Ejegg for the help with that.)

Reminder: For config changes on the main production wikis, please do not +2 (unless you are deploying). Thanks much!!

Closing as superseded by T232964.

I just double-checked this on the beta cluster, and it indeed seems to not be displaying protection actions after the initial banner creation.

Let's remove this page, as it seems it's currently not used. Please see T232912.

I've verified that both types of event are making it correctly from clients, through the Kafka streams and into log files in alnitak:/srv/banner_logs/. So, closing this task for now! Thanks much!!

Closing, since it seems this issue is settled Please re-open if I'm wrong... Thanks

Closing this task as plans for banner data collection have changed direction; we'll be using EventLogging for this in the near future. (See T183978.)

@Umherirrender, could you or anyone else perhaps confirm if this is still an issue today? Thanks much!!!

This has been working for a while, it just seems that we forgot to close it!

Verified that this is fixed!

This is included in 1.34.0-wmf.22, but because of a train delay, it hasn't been deployed to the group 2 (which includes Wikipedia). However, I've verified that it's working on other production wikis.

This is included in 1.34.0-wmf.22, but because of a train delay, it hasn't been deployed to the group 2 (which includes Wikipedia). However, I've verified that it's working on other production wikis.

Thanks for calling out!!!

After looking over some of the code, I guess the two possible causes seem to be (1) reading from a replica instead of master, or (2) stale object cache. Could be the same cause as T232793.

After looking over some of the code, I guess the two possible causes seem to be (1) reading from a replica instead of master, or (2) stale object cache. Could be the same cause as T232430.

For current discussion of possible causes, please see T232430. Thanks!!!

Ooops! It looks like the attached change targeted a different bug than the one described in this change... And while it fixed the issue for me locally, it's still occurring intermittently on production and on the Beta cluster.

The timeout issue seems to be fine now... Just confirming that we can close this? Thanks much!!!

In T232655#5485368, @Pcoombe wrote:

Sorry, this one was my fault. I recently added code in CoreJS to block typing letters in the "other monthly amount" field, which was a bug one of the QA specialist candidates spotted. But this throws an error if that field doesn't exist in the banner, i.e. before the introduction of the upsell. The much older banners were still working as they don't include this CoreJS.
Fixed it by using jQuery instead, which doesn't fail if the element doesn't exist.

Strangely, at least for me, the issue seems to have cleared itself up. All the links listed under "Links that don't work" are working fine for me as of this writing. However, I can also confirm that earlier today, stuff didn't work, including this link that was shared on IRC, which also works now: https://en.wikipedia.org/wiki/NASA?banner=trilogy_enWW_dsk_p1_lg_template_RML_upsell4&force=1

In T226655#5485083, @mepps wrote:

Should this be closed @AndyRussG? 3/4 of these are merged.

Done! Now deployed on production (along with T225101, T226967 and T229126).

I think the timeout issue should be solved by the attached change, just merged now.