Chrome bug makes us miss collecting Navigation Timing metrics for redirects.
Closed, ResolvedPublic
Actions

Assigned To

Authored By

	Peter
	Feb 21 2018, 8:00 AM

Description

There's a bug in Chrome (at least on Mac OS X) where redirects sets a wrong requestStart value for the Navigation Timing API. The value is lower than fetchStart (but should be higher) and we filter out those accesses with https://github.com/wikimedia/mediawiki-extensions-NavigationTiming/blob/master/modules/ext.navigationTiming.js#L59.

That means that we miss out of metrics when that happen when a user:

On https://www.wikipedia.org/ and click on one of the country Wikipedias: https://www.wikipedia.org -> https://en.wikipedia.org/ -> https://en.wikipedia.org/wiki/Main_Page

Users that internally search and we redirect the user to the the page: Search for Obama in the search field and click return: https://en.wikipedia.org/wiki/Main_Page -> https://en.wikipedia.org/w/index.php?search=Obama&title=Special%3ASearch&go=Go -> https://en.wikipedia.org/wiki/Barack_Obama

Upstream: https://bugs.chromium.org/p/chromium/issues/detail?id=813889

Related Objects

Mentioned In: T190977: Investigate NavigationTiming.nonCompliant error increase
Mentioned Here: T197954: Mobile Safari (WebView) produces invalid Navigation Timing data

Event Timeline

Peter created this task.Feb 21 2018, 8:00 AM

Restricted Application added a subscriber: Aklapper. · View Herald TranscriptFeb 21 2018, 8:00 AM

Peter added a subscriber: Krinkle.Feb 21 2018, 8:01 AM

Krinkle triaged this task as Medium priority.Feb 26 2018, 9:15 PM

Krinkle added a project: NavigationTiming.

• Imarlier moved this task from Inbox, needs triage to Blocked (old) on the Performance-Team board.Feb 26 2018, 9:21 PM

Krinkle changed the task status from Open to Stalled.Mar 5 2018, 9:26 PM

Krinkle moved this task from Backlog to Reported Upstream on the Upstream board.

Krinkle mentioned this in T190977: Investigate NavigationTiming.nonCompliant error increase.Mar 29 2018, 9:22 PM

Krinkle moved this task from Reported Upstream to Patch merged upstream on the Upstream board.Mar 29 2018, 9:27 PM

Seems to be fixed upstream, and I've confirmed the bug does not occur on a fresh https://en.wikipedia.org page view (which redirects to Main_Page), using Chrome canary 68.0.3409.0 (Mac).

The earliest chromium/src tag containing the fix is 67.0.3383.0, so looks like it'll make it in Chrome 67 rather than 68, which is already in Beta and should reach Stable on May 29.

Tested it manually in 67 (67.0.3396.79) and it seems to work correct now. Lets when we can see thing in the RUM metrics.

Need to check on what's driving error rate now.

• Imarlier moved this task from Blocked (old) to Doing (old) on the Performance-Team board.Jun 21 2018, 11:05 AM

Dump from navtiming.py logs on webperf1001 (last few hours, lightly normalised and sorted by frequency)

Rank	Hits	Normalised message
#1	1987	(is_compliant:443) Discarding event because responseEnd is out of order [{'os_minor': '4', 'os_major': '11', 'device_family': 'iPhone', 'os_family': 'iOS', 'browser_minor': '0', 'browser_major': '11', 'browser_family': 'Mobile Safari'}]
#2	1566	(is_compliant:443) Discarding event because requestStart is out of order [{'os_minor': '3', 'os_major': '9', 'device_family': 'iPad', 'os_family': 'iOS', 'browser_minor': '0', 'browser_major': '9', 'browser_family': 'Mobile Safari'}]
#3	821	(is_compliant:443) Discarding event because responseEnd is out of order [{'os_minor': '3', 'os_major': '11', 'device_family': 'iPhone', 'os_family': 'iOS', 'browser_minor': '0', 'browser_major': '11', 'browser_family': 'Mobile Safari'}]
#4	786	(is_compliant:443) Discarding event because requestStart is out of order [{'os_minor': '3', 'os_major': '9', 'device_family': 'iPhone', 'os_family': 'iOS', 'browser_minor': '0', 'browser_major': '9', 'browser_family': 'Mobile Safari'}]
#5	468	(is_compliant:443) Discarding event because responseEnd is out of order [{'os_minor': '2', 'os_major': '11', 'device_family': 'iPhone', 'os_family': 'iOS', 'browser_minor': '0', 'browser_major': '11', 'browser_family': 'Mobile Safari'}]

Marking this as resolved given the Chrome errors are no longer common. I've created T197954 for the well-known issue with Safari given there was no task for it yet.

Chrome bug makes us miss collecting Navigation Timing metrics for redirects.Closed, ResolvedPublicActions

Description

Related Objects

Event Timeline

Chrome bug makes us miss collecting Navigation Timing metrics for redirects.
Closed, ResolvedPublic
Actions