Fix Histogram Bad Construction Arguments |
|||||
Issue descriptionChrome Version: HEAD OS: ALL Some histograms have bad construction arguments. These are adjusted at run-time to work but are not necessarily correct. https://uma.googleplex.com/p/chrome/histograms/?endDate=latest&dayCount=28&histograms=Histogram.BadConstructionArguments&fixupData=true&showMax=true&filters=channel%2Ceq%2C1%2Cisofficial%2Ceq%2CTrue&implicitFilters=isofficial Login.ConsumerNewUsersAllowed: d32=-2055973512 Network.Shill.Ethernet.ExpiredLeaseLengthSeconds: d32=662206917 Network.Shill.Wifi.ExpiredLeaseLengthSeconds: d32=-1286300074 Network.Shill.Wifi.LinkMonitorBroadcastErrorsAtFailure: d32=1314934615 Network.Shill.Wifi.LinkMonitorResponseTimeSample: d32=424952287 Network.Shill.Wifi.LinkMonitorUnicastErrorsAtFailure: d32=825479942
,
Jun 29 2017
All of these seem to be part of some outside source, typically something related to Android. I think part of it is badly calculated boundary values: https://android.googlesource.com/platform/system/core.git/+/brillo-m7-release/metricsd/serialization/metric_sample.cc I don't really understand how these are coming into Chrome, though.
,
Jun 29 2017
They're part of the ChromeOS system. On ChromeOS, other processes than Chrome can use DBUS to communicate with Chrome to log histograms. Maybe just cc the histograms owners and have them fix the individual metrics?
,
Jun 29 2017
Got it. https://bugs.chromium.org/p/chromium/issues/detail?id=738122 I think the rest are all due to min=0 settings. Is that worth correcting? https://chromium.googlesource.com/chromiumos/platform/shill/+/master/metrics.cc const int Metrics::kMetricSignalAtDisconnectMin = 0; const int Metrics::kMetricNetworkSignalStrengthMin = 0; const int Metrics::kMetricLinkMonitorResponseTimeSampleMin = 0; const int Metrics::kMetricLinkMonitorSecondsToFailureMin = 0; const int Metrics::kMetricLinkMonitorErrorCountMin = 0; const int Metrics::kMetricCellularAutoConnectTotalTimeMin = 0; const int Metrics::kMetricCellularSignalStrengthBeforeDropMin = 0;
,
Jun 29 2017
Oh, wait. min=0 is silently changed to min=1. Need to look at those again.
,
Jun 29 2017
SimpleCache.App.IndexWriteInterval.Foreground: Compositing.Renderer.NumActiveLayers: Omnibox.SuggestionUsed.URL.Experimental.NavigationToFirstMeaningfulPaint: These are all constructed with constant arguments so should either be always wrong or never wrong. For now, assuming some kind of corruption, possibly while reading a persistent file.
,
Jun 30 2017
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/a78cd305087eda00c7cd274fba6f02839a71fc80 commit a78cd305087eda00c7cd274fba6f02839a71fc80 Author: bcwhite <bcwhite@chromium.org> Date: Fri Jun 30 14:47:26 2017 Enumerate some known-bad CromeOS histograms. BUG=736416 Review-Url: https://codereview.chromium.org/2970433002 Cr-Commit-Position: refs/heads/master@{#483703} [modify] https://crrev.com/a78cd305087eda00c7cd274fba6f02839a71fc80/tools/metrics/histograms/enums.xml
,
Nov 3 2017
,
Nov 6 2017
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/5b1fc679a9efd4909dd2f714b108c73a8465bac6 commit 5b1fc679a9efd4909dd2f714b108c73a8465bac6 Author: Brian White <bcwhite@chromium.org> Date: Mon Nov 06 18:26:48 2017 More histogram name hashes for those with bad construction. Bug: 736416 Change-Id: I3d98e6a1ebf0328a7cd22e9243ba678723b379a2 Reviewed-on: https://chromium-review.googlesource.com/753926 Reviewed-by: Alexei Svitkine <asvitkine@chromium.org> Commit-Queue: Brian White <bcwhite@chromium.org> Cr-Commit-Position: refs/heads/master@{#514186} [modify] https://crrev.com/5b1fc679a9efd4909dd2f714b108c73a8465bac6/tools/metrics/histograms/enums.xml
,
Nov 6 2017
,
Nov 6 2017
,
Jun 29 2018
|
|||||
►
Sign in to add a comment |
|||||
Comment 1 by bcwh...@chromium.org
, Jun 29 2017