Chromium Code Reviews
|
|
Chromium Code Reviews
Issue
2094453003:
Enable UMA tracking of --enable-experimental-web-platform-features (Closed)
|
Created:
4 years, 6 months ago by Rick Byers Modified:
4 years, 6 months ago Reviewers:
rkaplow CC:
chromium-reviews, asvitkine+watch_chromium.org Base URL:
https://chromium.googlesource.com/chromium/src.git@master Target Ref:
refs/pending/heads/master Project:
chromium Visibility:
Public. |
DescriptionEnable UMA tracking of --enable-experimental-web-platform-features
BUG=622495
Committed: https://crrev.com/dfc3cc7404ed5e3fb5eb50ffc731c75445497dc3
Cr-Commit-Position: refs/heads/master@{#401608}
Patch Set 1 #
Messages
Total messages: 16 (5 generated)
Description was changed from ========== Enable UMA tracking of --enable-experimental-web-platform-features BUG=622495 ========== to ========== Enable UMA tracking of --enable-experimental-web-platform-features BUG=622495 ==========
rbyers@chromium.org changed reviewers: + rkaplow@chromium.org
rkaplow@ PTAL at this little tweak Any known issues with this data (besides it being apparently broken on Android - crbug.com/622496)? I assume UserActions are a sample subset of total user activity (the absolute values seem too small to me). Seems like maybe a histogram would be better in the case of chrome://flags entries.
On 2016/06/22 23:41:03, Rick Byers wrote: > rkaplow@ PTAL at this little tweak > > Any known issues with this data (besides it being apparently broken on Android - > crbug.com/622496)? I assume UserActions are a sample subset of total user > activity (the absolute values seem too small to me). Seems like maybe a > histogram would be better in the case of chrome://flags entries. Both are supposed to be added. This flag is already tracked under the Login.CustomFlags histogram, see https://codesearch.chromium.org/chromium/src/tools/metrics/histograms/histogr... UserActions capture timestamps where histograms do not - so people use them to understand usage flows in Chrome. It's honestly not super critical for flags - in general using the flags histogram is probably just easier.
lgtm
On 2016/06/23 04:26:27, rkaplow wrote: > On 2016/06/22 23:41:03, Rick Byers wrote: > > rkaplow@ PTAL at this little tweak > > > > Any known issues with this data (besides it being apparently broken on Android > - > > crbug.com/622496)? I assume UserActions are a sample subset of total user > > activity (the absolute values seem too small to me). Seems like maybe a > > histogram would be better in the case of chrome://flags entries. > > Both are supposed to be added. This flag is already tracked under the > Login.CustomFlags histogram, see > https://codesearch.chromium.org/chromium/src/tools/metrics/histograms/histogr... Login.CustomFlags is ChromeOS specific (hence "Login"), right? I'd love to bring that to other platforms. > UserActions capture timestamps where histograms do not - so people use them to > understand usage flows in Chrome. It's honestly not super critical for flags - > in general using the flags histogram is probably just easier.
On 2016/06/23 15:02:49, Rick Byers wrote: > On 2016/06/23 04:26:27, rkaplow wrote: > > On 2016/06/22 23:41:03, Rick Byers wrote: > > > rkaplow@ PTAL at this little tweak > > > > > > Any known issues with this data (besides it being apparently broken on > Android > > - > > > crbug.com/622496)? I assume UserActions are a sample subset of total user > > > activity (the absolute values seem too small to me). Seems like maybe a > > > histogram would be better in the case of chrome://flags entries. > > > > Both are supposed to be added. This flag is already tracked under the > > Login.CustomFlags histogram, see > > > https://codesearch.chromium.org/chromium/src/tools/metrics/histograms/histogr... > > Login.CustomFlags is ChromeOS specific (hence "Login"), right? I'd love to > bring that to other platforms. Also I believe that code is executed only when you're restarting in order to apply new flags. Really what I think we want is how often (eg. on startup) all the flags are set. > > UserActions capture timestamps where histograms do not - so people use them to > > understand usage flows in Chrome. It's honestly not super critical for flags - > > in general using the flags histogram is probably just easier.
The CQ bit was checked by rbyers@chromium.org
CQ is trying da patch. Follow status at https://chromium-cq-status.appspot.com/patch-status/2094453003/1
On 2016/06/23 15:02:49, Rick Byers wrote: > On 2016/06/23 04:26:27, rkaplow wrote: > > On 2016/06/22 23:41:03, Rick Byers wrote: > > > rkaplow@ PTAL at this little tweak > > > > > > Any known issues with this data (besides it being apparently broken on > Android > > - > > > crbug.com/622496)? I assume UserActions are a sample subset of total user > > > activity (the absolute values seem too small to me). Seems like maybe a > > > histogram would be better in the case of chrome://flags entries. > > > > Both are supposed to be added. This flag is already tracked under the > > Login.CustomFlags histogram, see > > > https://codesearch.chromium.org/chromium/src/tools/metrics/histograms/histogr... > > Login.CustomFlags is ChromeOS specific (hence "Login"), right? I'd love to > bring that to other platforms. > > > UserActions capture timestamps where histograms do not - so people use them to > > understand usage flows in Chrome. It's honestly not super critical for flags - > > in general using the flags histogram is probably just easier. Oh, you're correct - I actually didn't realize that. That does seems a bit silly. At any rate, you can use the UserActions.Counts psuedohistogram to track this and other flags once they have actions associated. Although honestly this might all be simpler just as a histogram and dropping actions entirely for flags.
On 2016/06/23 15:05:50, rkaplow wrote: > On 2016/06/23 15:02:49, Rick Byers wrote: > > On 2016/06/23 04:26:27, rkaplow wrote: > > > On 2016/06/22 23:41:03, Rick Byers wrote: > > > > rkaplow@ PTAL at this little tweak > > > > > > > > Any known issues with this data (besides it being apparently broken on > > Android > > > - > > > > crbug.com/622496)? I assume UserActions are a sample subset of total user > > > > activity (the absolute values seem too small to me). Seems like maybe a > > > > histogram would be better in the case of chrome://flags entries. > > > > > > Both are supposed to be added. This flag is already tracked under the > > > Login.CustomFlags histogram, see > > > > > > https://codesearch.chromium.org/chromium/src/tools/metrics/histograms/histogr... > > > > Login.CustomFlags is ChromeOS specific (hence "Login"), right? I'd love to > > bring that to other platforms. > > > > > UserActions capture timestamps where histograms do not - so people use them > to > > > understand usage flows in Chrome. It's honestly not super critical for flags > - > > > in general using the flags histogram is probably just easier. > > Oh, you're correct - I actually didn't realize that. That does seems a bit > silly. > > At any rate, you can use the UserActions.Counts psuedohistogram to track this > and other flags once they have actions associated. Although honestly this might > all be simpler just as a histogram and dropping actions entirely for flags. Yeah, even checking UserActions.Counts only looks at when the flag was flipped - not the flag settings (on restart).
Message was sent while issue was closed.
Description was changed from ========== Enable UMA tracking of --enable-experimental-web-platform-features BUG=622495 ========== to ========== Enable UMA tracking of --enable-experimental-web-platform-features BUG=622495 ==========
Message was sent while issue was closed.
Committed patchset #1 (id:1)
Message was sent while issue was closed.
Description was changed from ========== Enable UMA tracking of --enable-experimental-web-platform-features BUG=622495 ========== to ========== Enable UMA tracking of --enable-experimental-web-platform-features BUG=622495 Committed: https://crrev.com/dfc3cc7404ed5e3fb5eb50ffc731c75445497dc3 Cr-Commit-Position: refs/heads/master@{#401608} ==========
Message was sent while issue was closed.
Patchset 1 (id:??) landed as https://crrev.com/dfc3cc7404ed5e3fb5eb50ffc731c75445497dc3 Cr-Commit-Position: refs/heads/master@{#401608}
Message was sent while issue was closed.
On 2016/06/23 15:05:50, rkaplow wrote: > On 2016/06/23 15:02:49, Rick Byers wrote: > > On 2016/06/23 04:26:27, rkaplow wrote: > > > On 2016/06/22 23:41:03, Rick Byers wrote: > > > > rkaplow@ PTAL at this little tweak > > > > > > > > Any known issues with this data (besides it being apparently broken on > > Android > > > - > > > > crbug.com/622496)? I assume UserActions are a sample subset of total user > > > > activity (the absolute values seem too small to me). Seems like maybe a > > > > histogram would be better in the case of chrome://flags entries. > > > > > > Both are supposed to be added. This flag is already tracked under the > > > Login.CustomFlags histogram, see > > > > > > https://codesearch.chromium.org/chromium/src/tools/metrics/histograms/histogr... > > > > Login.CustomFlags is ChromeOS specific (hence "Login"), right? I'd love to > > bring that to other platforms. > > > > > UserActions capture timestamps where histograms do not - so people use them > to > > > understand usage flows in Chrome. It's honestly not super critical for flags > - > > > in general using the flags histogram is probably just easier. > > Oh, you're correct - I actually didn't realize that. That does seems a bit > silly. > > At any rate, you can use the UserActions.Counts psuedohistogram to track this > and other flags once they have actions associated. Although honestly this might > all be simpler just as a histogram and dropping actions entirely for flags. Yeah, filed https://crbug.com/622729 to track. I probably don't have time to dig into this right now, but might be the better way of trying to fix the issue on Android so maybe I will sometime soon... |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
