Investigate SubresourceFilter.IndexRuleset.Verify2.WallDuration regression |
||
Issue descriptionThere looks like a regression in this metric after we turned on the AdTagging experiment. https://uma.googleplex.com/p/chrome/timeline_v2/?sid=e86402e00caa62f68ac549551c8ee6a3 Unfortunately we lost data for a few days due to bad finch config timing. However, it does seem like the regression must be in this range: https://chromium.googlesource.com/chromium/src/+log/66.0.3343.3..66.0.3350.0?pretty=fuller&n=10000 I don't see anything very suspicious, and total counts are pretty similar. In other news. Wow. Ruleset verification takes a _long_ time on Windows. We might not want to block the first navigation waiting for this to complete for ad tagging.
,
Mar 13 2018
The one thing that changed in the field trial config for Ad Tagging is that it is per-session, not permanent. Perhaps that is the cause?
,
Mar 26 2018
r542132 also caused a regression at high %iles, likely due to startup being a bit busier.
,
Mar 26 2018
Interesting, there is a good bit less volume on the 23rd, which might be a cause.
,
Jun 19 2018
Archiving. I don't think there's anything we can do here. |
||
►
Sign in to add a comment |
||
Comment 1 by csharrison@chromium.org
, Mar 13 2018