🐛 Use timeOrigin for getFirstVisibleTime when page starts visible #36273
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
samouri
reviewed
Oct 7, 2021
There was a problem hiding this comment.
Thank you @jridgewell for tracking this down!
Fixing visible time calculation within Ampdoc provides a more robust fix for #34383. We can also slightly simplify the implementation of tickSinceVisible_ now, by removing the conditional around visibleTime calculation.
samouri
approved these changes
Oct 7, 2021
jridgewell
force-pushed
the
getFirstVisibleTime
branch
from
61008c3
to
58627a7
Compare
jridgewell
force-pushed
the
getFirstVisibleTime
branch
from
58627a7
to
1ef1d1a
Compare
Merged
estherkim
pushed a commit
to estherkim/amphtml
that referenced
this pull request
Nov 17, 2021
…pproject#36273) * Use timeOrigin for getFirstVisibleTime when page starts visible * Update old tests * Add tests for firstVisibleTime * Fix global conflict * Fix tests
jridgewell
added a commit
to jridgewell/amphtml
that referenced
this pull request
Nov 22, 2021
Another fix for ampproject#36273
Merged
jridgewell
added a commit
that referenced
this pull request
Nov 22, 2021
* Fix DocInfo tests Another fix for #36273 * Remove debugger statement * Lint
jridgewell
added a commit
to jridgewell/amphtml
that referenced
this pull request
Dec 7, 2021
In ampproject#36273, I switched from `Date.now()` to `performance.now()`, thinking it was a higher resolution timestamp time since the unix epoch. But, `peformance.now` is relative to _page load time_ (or `timeOrigin`), not the epoch. This matters because our "since visible" metrics use the epoch as their base time, meaning the timestamps are roughly 1.6 trillion milliseconds. But `performance.now` is tiny, in the range of a few million for a page that was loaded an hour ago. When we calculated the "since visible", we received values that were still in the trillions when we expect millions. By adding `performance.now()` and `performance.timeOrigin`, we get a timestamp relative to the epoch. Now when we calculate the "since visible", we get the correct values.
This was referenced Dec 7, 2021
jridgewell
added a commit
to jridgewell/amphtml
that referenced
this pull request
Dec 7, 2021
In ampproject#36273, I switched from `Date.now()` to `performance.now()`, thinking it was a higher resolution timestamp time since the unix epoch. But, `peformance.now` is relative to _page load time_ (or `timeOrigin`), not the epoch. This matters because our "since visible" metrics use the epoch as their base time, meaning the timestamps are roughly 1.6 trillion milliseconds. But `performance.now` is tiny, in the range of a few million for a page that was loaded an hour ago. When we calculated the "since visible", we received values that were still in the trillions when we expect millions. By adding `performance.now()` and `performance.timeOrigin`, we get a timestamp relative to the epoch. Now when we calculate the "since visible", we get the correct values.
jridgewell
added a commit
that referenced
this pull request
Dec 7, 2021
In #36273, I switched from `Date.now()` to `performance.now()`, thinking it was a higher resolution timestamp time since the unix epoch. But, `performance.now` is relative to _page load time_ (or `performance.timeOrigin`), not the epoch. This matters because our "since visible" metrics use the epoch as their base time, meaning the timestamps are roughly 1.6 trillion milliseconds. But `performance.now` is tiny, in the range of a few million for a page that was loaded an hour ago. When we calculated the "since visible", we received values that were still in the trillions when we expect millions. By adding `performance.now()` and `performance.timeOrigin`, we get a timestamp relative to the epoch. Now when we calculate the "since visible", we get the correct values.
ampprojectbot
pushed a commit
that referenced
this pull request
Dec 8, 2021
In #36273, I switched from `Date.now()` to `performance.now()`, thinking it was a higher resolution timestamp time since the unix epoch. But, `performance.now` is relative to _page load time_ (or `performance.timeOrigin`), not the epoch. This matters because our "since visible" metrics use the epoch as their base time, meaning the timestamps are roughly 1.6 trillion milliseconds. But `performance.now` is tiny, in the range of a few million for a page that was loaded an hour ago. When we calculated the "since visible", we received values that were still in the trillions when we expect millions. By adding `performance.now()` and `performance.timeOrigin`, we get a timestamp relative to the epoch. Now when we calculate the "since visible", we get the correct values. (cherry picked from commit 7645c4e)
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
When comparing the LCP and LCPV scores for viewer-embedded pages that are immediately visible, I noticed the LCPV scores were considerably lower. This seems wrong, if the page is immediately visible, the LCP and LCPV should be equivalent.
I think this is because the
ampdockeeps track of the time the page is first visible, but that's initialized in JS when theampdocis first created. That means we don't notice the parse+exec time that it takes to reach that point, and the LCPV is better than reality.