Tue, Feb 18
Fri, Feb 14
The second one duplicates T226728.
Seems like there are two issues here:
Wed, Feb 12
Mon, Feb 10
Sat, Feb 8
Fri, Feb 7
Thu, Feb 6
Wed, Feb 5
These are merged:
Tue, Feb 4
wmf.16 deployed to all wikis at 2020-02-03 23:24 UTC.
We're definitely not treating this as a true blocker, given that wmf.16 is fully deployed at this point.
After discussion with @thcipriani and combing over graphs / logs from earlier, it started to look like the spike here might be purely coincidental. Things seem essentially stable since a redeploy at 23:24:52. I'll monitor for the next few hours and treat this as resolved if it doesn't flare back up.
Mon, Feb 3
Rolling to group2 appeared to coincide again with a spike in response times and, this time, exceptions/fatals:
FWIW, a note that I'm seeing breakage for new runs:
I assume that this issue didn't occur when rolling out to any other group? i.e., there was no noticeable spike in CPU usage after group1?
Fri, Jan 24
There are no deployments due to All Hands next week, right?
@D3r1ck01 - just for clarity, are you still proposing 565762: PathRouter: Allow duplicate names for subpatterns in RegEx as a fix for this one?