yoangela@chromium.org chrome-omnibox-team@google.com Records number of accepted keyword searches enumerated by how the user entered the keyword mode. See also Omnibox.EnteredKeywordMode2 for the total number of times a user enters keyword mode whether or not the keyword suggestion was accepted. Note that this only records when a user accepts a search suggestion for the keyword search engine and does not record for URL suggestions even if the user is in keyword mode. ender@google.com chrome-mobile-search@google.com For action of type "{Type}", captures whether the user interacted with this action to complete the interaction with the omnibox. The histogram is captured only if the action "{Type}" was presented to the user when the interaction with the omnibox completed. The sum of "Used" and "Not Used" indicates the sum of all impressions of a given action. Action in Suggest are the Omnibox Action chips attached to a search suggestion. ender@google.com chrome-mobile-search@google.com Counts how often Omnibox Action in Suggest are {ShownOrUsed} and records the type (Omnibox Action in Suggest Type) of every {ShownOrUsed} action. Recorded once at the end of the interaction with the Omnibox Suggestions for every Action presented to the user. Not recorded for Suggestions that are annotated with Action in Suggest, where the Action in Suggest row was not shown to the user. Action in Suggest are the Omnibox Action chips attached to a search suggestion. jdonnelly@chromium.org chrome-omnibox-team@google.com For each answer received in suggest responses, the number that are well-formed and contain all the required elements. jdonnelly@chromium.org chrome-omnibox-team@google.com The number of times each omnibox suggestion answer type (e.g., weather, sports score) was received and parsed successfully. Can be normalized with the count of emits to Omnibox.SuggestRequest.Success.GoogleResponseTime, which counts all successful suggest responses from Google, not just those with answers attached. Note that GoogleResponseTime is technically off given that it will still count cases where we fail to extract, deserialize, or parse the response. But these cases are rare to non-existent. mahmadi@chromium.org chrome-omnibox-team@google.com Records the length of the aqs= param in the Search Result Page request URLs. Recorded every time a Search Result Page URL is generated in the omnibox/realbox. This happens either when the user selects a suggestion in order to load Search Result Page or when a proactive request is made to prefetch the Search Result Page. manukh@chromium.org jdonnelly@chromium.org chrome-omnibox-team@google.com When the user focuses or types into the omnibox, the dropdown results can be updated multiple times per input (i.e. when focused, per char typed, backspace, etc). Specifically, the set of events that trigger dropdown updates are: - synchronous suggestions - asynchronous suggestions. Can trigger multiple times per input. These are often generated on different threads. - expiring copied matches - interruptions e.g. additional user input changes or closing the dropdown A sample flow looks like: 1) User types a character. 2) 10ms later, update results with synchronous suggestions. 3) 50ms later, update results with some asynchronous suggestions. 4) 50ms later, update results with more asynchronous suggestions. 5) 400ms later, update results by removing copied matches from the previous input. 6) 50 later, update results with the last batch of asynchronous suggestions. The 'Provider.*' slices measure the time between the input change and the last async update of that provider is done or interrupted. Includes async updates that have no user visible change (i.e. that do not produce async matches or produce only async matches that get deduped, culled, or are the same as the existing match in its position). Only logged for asynchronous providers. Emitted exactly once per input per asynchronous provider; additionally emits to exactly one of the 'Completed' and 'Interrupted' slices per input per asynchronous provider. {Completed} manukh@chromium.org jdonnelly@chromium.org chrome-omnibox-team@google.com When the user focuses or types into the omnibox, the dropdown results can be updated multiple times per input (i.e. when focused, per char typed, backspace, etc). Specifically, the set of events that trigger dropdown updates are: - synchronous suggestions - asynchronous suggestions. Can trigger multiple times per input. These are often generated on different threads. - expiring copied matches - interruptions e.g. additional user input changes or closing the dropdown A sample flow looks like: 1) User types a character. 2) 10ms later, update results with synchronous suggestions. 3) 50ms later, update results with some asynchronous suggestions. 4) 50ms later, update results with more asynchronous suggestions. 5) 400ms later, update results by removing copied matches from the previous input. 6) 50 later, update results with the last batch of asynchronous suggestions. {Change} {Completed} manukh@chromium.org jdonnelly@chromium.org chrome-omnibox-team@google.com The time `AutocompleteController::UpdateResult()` took. That method is invoked, and this histogram is recorded: a) Once for the sync autocomplete update that occurs when the omnibox input changes (e.g. the user types a character, pastes text, focuses the omnibox, etc). b) Once for each async autocomplete update that occurs asyncly after the omnibox input changes and the various async providers complete. Usually occurs up to 3 times per input change. But can be less if the updates are interrupted (e.g. the popup is closed or input changed again) or include no new suggestions. c) Once when copied matches from the previous autocomplete pass expire. This happens at most once per input change, but may not occur if all providers complete or the input changes prior to the expiration. Because this encompasses both sync and async omnibox updates, and the number of async updates per sync update is not fixed, it's not a good representative of how a feature effects user experienced latency. E.g., a feature that makes users type slower will bias this metric towards async updates. Besides, this is only a small segment of the user experienced latency. This is only intended to measure changes in the actual implementation of `UpdateResult()`. {Slice} This metric uses SCOPED_UMA_HISTOGRAM_TIMER_MICROS so clients without high-resolution clocks will either a) be dropped, b) report 0 for very short times, or c) be recorded as expected (depending on which comments you believe ¯\_(ツ)_/¯). 3/2022; feature launched. manukh@chromium.org jdonnelly@chromium.org chrome-omnibox-team@google.com The time elapsed between the ChromeOmniboxClient requesting a suggestion's image to receiving the image. This includes both uncached and cached requests. See BitmapFetcher.Fetch and BitmapFetcher.Decode for what contributes to uncached images latency; however, the sum of those metrics can be greater than this metric as image prefetching begins before the ChromeOmniboxClient requests an image. peilinwang@google.com woa-performance-bugs+jank@google.com Measures the amount of time it takes to call UrlBar.calculateVisibleHint(). Recorded once per call. jdonnelly@chromium.org mpearson@chromium.org asvitkine@chromium.org Records the time taken between a keystroke being typed in the omnibox and the text being painted. If there are multiple keystrokes before a paint, logs the time since the earliest one. This duration is composed of three parts: a) the time spent processing the initial input event b) the time spent for the repaint task to be scheduled on the message loop c) the time spent painting the Omnibox d) (on views platforms) the time until the pixels are actually composited There's a number of breakdown metrics to help diagnose a regression. First, Omnibox.CharTypedToRepaintLatency.ToPaint measures the combined time of (a) and (b). Omnibox.QueryTime2 is a good proxy for just (a). And there's also Omnibox.PaintTime that corresponds to (c). We don't have a direct metric for (d), but if neither Omnibox.CharTypedToRepaintLatency.ToPaint nor Omnibox.PaintTime regressed, then the regression must be in (d). Note: The semantics of this metric on views platforms changed in M62, as previously time (d) was not included in the metric. asvitkine@chromium.org mpearson@chromium.org jdonnelly@chromium.org A refinement of Omnibox.CharTypedToRepaintLatency metric. It measures the time between the first character insertion in a series that happen during a frame, to the time the compositor reports the text was rendered on-screen. Compositor-provided timestamps are used for this version of the metric, so it is expected to be more accurate: the baseline CharTypedToRepaintLatency uses now() from when the subsequent has-processed notification arrives back on the UI thread, which may overestimate user-visible latency. Note: Compositor-provided timestamps are taken using presentation time callbacks. Before M110 the callbacks used for this histogram might be called even if the next submitted frame fails to present which means that the character is not really rendered on-screen yet. From M110 onward, this is fixed by using callbacks that are only called after the first following successful presentation, so a slight change might be noticed in the histogram values. asvitkine@chromium.org mpearson@chromium.org jdonnelly@chromium.org Records the time between when OnPaint() is called to the time the compositor reports pixels were successfully drawn to the screen. This a subset of the time reported in Omnibox.CharTypedToRepaintLatency.InsertToPresent. Note: Compositor-provided timestamps are taken using presentation time callbacks. Before M110 the callbacks used for this histogram might be called even if the next submitted frame fails to present which means that the character is not really rendered on-screen yet. From M110 onward, this is fixed by using callbacks that are only called after the first following successful presentation, so a slight change might be noticed in the histogram values. Warning: this histogram was expired from 2021-06-30 to 2022-12-12; data may be missing. asvitkine@chromium.org mpearson@chromium.org jdonnelly@chromium.org Records the time taken between a keystroke being typed in the omnibox and the time when we're ready to paint the omnibox. This is a breakdown diagnostic metric for Omnibox.CharTypedToRepaintLatency, to help diagnose regressions in that metric. See that metric's description for more info. gangwu@chromium.org jdonnelly@chromium.org Recorded every time the clipboard suggestion is removed from omnibox suggestion list and the system clipboard. The value indicates the estimated age of the clipboard. Note: recording gap between 2022/09 and 2023/08. gangwu@chromium.org jdonnelly@chromium.org Recorded every time the omnibox is focussed and recent content from the user's clipboard is suggested. The value indicates the estimated age of the clipboard. (If Chrome observed the clipboard modification, this age is exact. If Chrome did not observe the modification, then it's a conservative estimate: the last time Chrome observed a clipboard modification, which is certainly older than the current clipboard. If Chrome never observed a clipboard modification, no clipboard suggestion is shown, meaning this histogram will never be emitted to.) Intended to be compared with MobileOmnibox.PressedClipboardSuggestionAge. gangwu@chromium.org jdonnelly@chromium.org Recorded every time the omnibox is focussed and a recent content from the user's clipboard is suggested. The number emitted is the number of times the clipboard content has been suggested within the same session including the current time. Thus, the third time it is shown, we'll emit a three to this histogram, and this histogram will have previously seen emits of one and two. If the clipboard content was the same during a previous run of Chrome and this clipboard content was suggested during that run, those impressions are not counted. Also, if the clipboard content changes during a particular run of Chrome to other content, the omnibox is focused and that clipboard content is suggested, then content changes back and Chrome starts suggesting the older clipboard content again, the counts start again from scratch. Chrome only remembers the number of times the clipboard content was shown consecutively. This value is useful to compare with the number of times a clipboard suggestion has been shown when it is clicked. This value can be obtained from OmniboxEvent records in which the selected suggestion is from Clipboard provider. In those cases, look in the Clipboard provider's ProviderInfo field for |times_returned_results_in_session|. Note that at the time of this writing that OmniboxEvent logs aren't recorded in incognito whereas histograms are. Thus, the total counts will not be comparable, though the distributions should be. gangwu@chromium.org jdonnelly@chromium.org Recorded every time the omnibox is focussed and a recent content from the user's clipboard is suggested. The value indicates whether the current URL was shown (which would appear above the clipboard suggestion) or was absent (which ought to only happen when the omnibox is empty / the user is on the NTP). On Android, the total count for this histogram can be usefully compared with the count of the user action FocusLocation in order to determine the rate at which a clipboard suggestion is shown in the omnibox. To determine the same rate on iOS, one needs to compare this histogram's count with the sum of the user actions MobileFocusedOmniboxOnNtp, MobileFocusedFakeboxOnNtp, and MobileFocusedOmniboxNotOnNtp. Furthermore, on either platform, this histogram's count can be usefully compared with the count in the clipboard bucket of the Omnibox.SuggestionUsed.Provider histogram to determine the clickthrough rate on these suggestions. Note: recording gap between 2022/10 and 2023/08. jdonnelly@chromium.org mpearson@chromium.org chrome-omnibox-team@google.com The number of cut or copy commands on all selected text in the omnibox. Gathered on desktop platforms (Win, Mac, Linux, Chrome OS). manukh@chromium.org mpearson@chromium.org jdonnelly@chromium.org skare@chromium.org Counts the number of document suggest requests the omnibox sent, were invalidated, and were completed successfully. manukh@chromium.org jdonnelly@chromium.org Measures the duration between when the async backend request was sent and response received. Recorded exactly once per backend request. This will be a subset of times document suggestions were requested, as 1) the provider is debounced, 2) the provider may be interrupted (e.g. due to a new input) before the request is sent (i.e. before the oauth token is ready), and 3) some provider requests are filtered (e.g. input too short). manukh@chromium.org jdonnelly@chromium.org Measures the duration between when the async backend request was sent and response received. Recorded at most once per backend request. This will be a subset of times document suggestions were requested, as 1) the provider is debounced, 2) the provider may be interrupted (e.g. due to a new input) before the request is sent (i.e. before the oauth token is ready), and 3) some provider requests are filtered (e.g. input too short). Only recorded if the backend request was invalidated; i.e., the provider was interrupted before the response was received. manukh@chromium.org jdonnelly@chromium.org Measures the duration between when the async backend request was sent and response received. Recorded at most once per backend request. This will be a subset of times document suggestions were requested, as 1) the provider is debounced, 2) the provider may be interrupted (e.g. due to a new input) before the request is sent (i.e. before the oauth token is ready), and 3) some provider requests are filtered (e.g. input too short). Only recorded if the backend request was not invalidated; i.e., the provider was not interrupted before the response was received. manukh@chromium.org mpearson@chromium.org jdonnelly@chromium.org skare@chromium.org Number of results returned in each document suggestion reply. Logged for successful requests where the provider returned a parseable result set. manukh@chromium.org jdonnelly@chromium.org Measures the duration between when the provider begins (after a debouncing delay) and ends. Recorded exactly once per processed provider request. This will be a subset of times document suggestions were requested, as 1) the provider is debounced, and 2) some provider requests are filtered (e.g. input too short). manukh@chromium.org jdonnelly@chromium.org Measures the duration between when the provider begins (after a debouncing delay) and ends. Recorded at most once per processed provider request. This will be a subset of times document suggestions were requested, as 1) the provider is debounced, and 2) some provider requests are filtered (e.g. input too short). Only recorded if the provider was interrupted (e.g. due to a new input). manukh@chromium.org jdonnelly@chromium.org Measures the duration between when the provider begins (after a debouncing delay) and ends. Recorded at most once per processed provider request. This will be a subset of times document suggestions were requested, as 1) the provider is debounced, and 2) some provider requests are filtered (e.g. input too short). Only recorded if the provider was not interrupted (e.g. due to a new input). jdonnelly@chromium.org mpearson@chromium.org chrome-omnibox-team@google.com The new enumeration of how many times users enter keyword hint mode "Search ___ for:" and how. Note that we don't recognize the changing of the keyword itself as entering keyword mode, if the user never left keyword mode e.g. if the user was arrowing around the suggestions, so we'll preserve the original keyword mode entry method in this case. manukh@chromium.org jdonnelly@chromium.org chrome-omnibox-team@google.com When the omnibox is focused and escape is pressed, the 1st of the following occurs: 1) If there is temporary text (i.e. a non default suggestion is selected), it is cleared. 2) If the popup is open, the popup is closed. 3) If the popup is open, which is considered as user input in progress, then the popup is closed and user input is cleared. 4) If user input is in progress, it is cleared. This state (i.e. input in progress with the popup closed) typically occurs when input was in progress, then the omnibox was blurred, and then it was refocused. 5) The omnibox is blurred. This histogram records which of the above occurred. It is recorded every time escape is pressed with the omnibox focused. Pressing escape repeatedly e.g. 10 times will emits at most 4 times until the omnibox is blurred. ender@chromium.org jdonnelly@chromium.org mpearson@chromium.org tommycli@chromium.org This records whether user initiated navigation after focusing the Omnibox versus switches focus elsewhere without navigating. This metric includes accidental focus-then-losing-focus events, including brief ones. It also includes ones in which the user did not type any text or otherwise engage with the Omnibox. E.g., the user tabbing through all top Chrome UI controls will cause this metric to be recorded. If the user switches focus from the Chrome window to another window while the Omnibox is focussed, that also counts as a loss of focus. This metric is not recorded for the Realbox interactions. Recorded every time the Omnibox focus is cleared. jdonnelly@chromium.org mpearson@chromium.org chrome-omnibox-team@google.com The length of time between when a user focused on the omnibox and first modifies the omnibox. jdonnelly@chromium.org mpearson@chromium.org chrome-omnibox-team@google.com The length of time between when a user focused on the omnibox and opened an omnibox match (which could be what they typed or a suggestion). This is recorded regardless of whether the omnibox dropdown (a.k.a. popup) is open. It is not recorded if a match is opened without triggering a focus event, e.g., when a user drags a URL to the omnibox to navigate. To know how common this last condition is, compare the total count of this histogram to the total number of omnibox events. mahmadi@google.com chrome-omnibox-team@google.com Counts the number of times a suggestion group ID is prevented form appearing in the results. This is recorded when the user hides a set of grouped suggestions, such as trending zero-suggest, in the omnibox UI using the control in the grouped suggestions' header. mahmadi@google.com chrome-omnibox-team@google.com Counts the number of times a suggestion group ID is permitted to appear in the results. This is recorded when the user makes a set of previously hidden grouped suggestions, such as trending zero-suggest, visible in the omnibox UI using the control in the grouped suggestions' header. orinj@chromium.org jdonnelly-team@google.com Fuzzy match conversion produces modified matches from other synchronous autocomplete providers (sub-providers) using alternative inputs: what the user might have typed, the automatic corrections. This metric is logged once per sub-provider per omnibox input if and only if the fuzzy provider runs a search (not all omnibox inputs result in search) and finds any corrections (this includes the zero case where corrections are found but produce no matches). Logs for each fuzzy sub-provider the number of correction matches converted to fuzzy provider matches for a single omnibox input. This measures how many 'best' matches (among all the sub-provider's matches for a specific correction) are converted from this sub-provider, summed for the various corrections of a single user input (will not exceed `provider_max_matches_` even in unlikely extreme cases). This is useful in evaluating the efficacy of various sub-providers for fuzzy matching. orinj@chromium.org jdonnelly-team@google.com When a fuzzy match was offered at the moment the user opened a match, this measures whether a fuzzy suggestion was taken. This is only recorded when at least one fuzzy match is in the result set, so it is useful to measure total presence and engagement with fuzzy URL suggestions. It can be interpreted as a click through rate for fuzzy match presentation as a whole: given that any fuzzy matches were presented at the moment a match was selected by the user, how often was the selected match a fuzzy URL suggestion? Context: Fuzzy match conversion produces modified matches from other synchronous autocomplete providers (sub-providers) using alternative inputs: what the user might have typed, the automatic corrections. orinj@chromium.org jdonnelly-team@google.com The time taken by HistoryFuzzyProvider to search for fuzzy input corrections. This is measured if and only if a search is performed, so many input states will not result in recording this metric (for example if the input contains a space, that's taken as a hint that the user may be searching instead of typing a domain name so no fuzzy search is performed). It is useful to distinguish portions of the total time taken by the provider, particularly as separate from sub-provider match gathering, so that if there are any pathological cases in the search algorithm then they will be apparent here. Context: Fuzzy match conversion produces modified matches from other synchronous autocomplete providers (sub-providers) using alternative inputs: what the user might have typed, the automatic corrections. estark@chromium.org carlosil@chromium.org The time duration that the user hovers the mouse over the omnibox before exiting the hover or focusing the omnibox. Whenever the user's mouse enters the omnibox, this histogram is recorded when the mouse next exits or the omnibox is focused, whichever comes first. The recorded value is a delta from a timestamp noted when the user's mouse first entered the omnibox. jdonnelly@chromium.org mpearson@chromium.org chrome-omnibox-team@google.com The kind of input the user provided when using the omnibox to go somewhere. The type can be misleading. For example if the user typed 'http:/', it gets marked as a query because it cannot be opened as a URL even though the user probably wanted and selected a URL from the list of suggestions. In M-74 and earlier, this was only recorded if the user had no incognito windows open. In M-75 and later, it records all omnibox interactions regardless, in order to be consistent with how user actions and histograms are handled. gjc@chromium.org blink-network-stack@google.com The number of IPv4 address parts input by the user. The histogram is used to verify whether shortened IPv4 addresses like 127.1 are still used. This histogram is recorded when the user fully types the URL in the omnibox and the URL uses an IPv4 address or IPv4 embedded IPv6 address as the host. jdonnelly@chromium.org chrome-omnibox-team@google.com Whether an omnibox interaction is a paste-and-search/paste-and-go action. (This histogram records both of these in the "True" bucket for this histogram because both of these are referred to as paste-and-go in the code.) These typically involve right-clicking in the omnibox and selecting that option from the dropdown. In M-74 and earlier, this was only recorded if the user had no incognito windows open. In M-75 and later, it records all omnibox interactions regardless, in order to be consistent with how user actions and histograms are handled. jdonnelly@chromium.org chrome-omnibox-team@google.com Whether the omnibox popup (a.k.a. dropdown) is open at the time the user used the omnibox to go somewhere. It can be closed if, for instance, the user clicked in the omnibox and hit return to reload the same page. Also, because paste-and-search/paste-and-go actions ignore the current content of the omnibox dropdown (if it is open) when they happen, we pretend the dropdown is closed when logging these. In M-74 and earlier, this was only recorded if the user had no incognito windows open. In M-75 and later, it records all omnibox interactions regardless, in order to be consistent with how user actions and histograms are handled. jdonnelly@chromium.org mpearson@chromium.org chrome-omnibox-team@google.com Whether the user deleted text immediately before selecting an omnibox suggestion. This is usually the result of pressing backspace or delete. In M-74 and earlier, this was only recorded if the user had no incognito windows open. In M-75 and later, it records all omnibox interactions regardless, in order to be consistent with how user actions and histograms are handled. yoangela@chromium.org chrome-omnibox-team@google.com Records number of times users use keyword mode enumerated by the type of search engine (whether the engine is "built-in"). {Usage} Expired in M110. mahmadi@chromium.org chrome-omnibox-team@google.com The number of search terms extracted from the database to produce local prefix search suggestions. Emitted when prefix search suggestions are requested as a result of the user typing into the Omnibox or the Realbox on the NTP. Expired in M110 in favor of Omnibox.LocalHistoryPrefixSuggest.SearchTermsExtractionTimeV2 mahmadi@chromium.org chrome-omnibox-team@google.com The amount of time it takes to extract search terms from the database and sort them by recency (if needed) to produce local prefix search suggestions. Emitted when prefix search suggestions are requested as a result of the user typing into the Omnibox or the Realbox on the NTP. mahmadi@chromium.org chrome-desktop-search@google.com The amount of time it takes to extract up to the requested number of search terms from the database for both the default and the keyword providers (if applicable) and sort them by recency to produce local prefix search suggestions. Emitted when prefix search suggestions are requested as a result of the user typing into the Omnibox or the NTP Realbox. mahmadi@chromium.org chrome-omnibox-team@google.com The length of time it takes for the corresponding URLs that would produce a deleted match to be queried asynchronously and deleted. Emitted when user deletes a local history zero-prefix suggestion. Expired in M110. mahmadi@chromium.org chrome-omnibox-team@google.com The number of search terms extracted from the database to produce local zero-prefix suggestions. Emitted when zero-prefix suggestions are requested as a result of the user focusing into the Omnibox or the Realbox on the NTP. Expired in M110 in favor of Omnibox.LocalHistoryZeroSuggest.SearchTermsExtractionTimeV2 mahmadi@chromium.org chrome-omnibox-team@google.com The amount of time it takes to extract search terms from the database and sort them by frecency to produce local zero-prefix suggestions. Emitted when zero-prefix suggestions are requested as a result of the user focusing into the Omnibox or the Realbox on the NTP. In M103, this was changed to no longer include the construction of the AutocompleteMatch values. mahmadi@chromium.org chrome-desktop-search@google.com The amount of time it takes to extract up to the requested number of search terms from the database and sort them by frecency to produce local zero-prefix suggestions. Emitted when zero-prefix suggestions are requested as a result of the user focusing into the Omnibox or the NTP Realbox. mahmadi@chromium.org chrome-omnibox-team@google.com The length of time it takes to handle the deletion request for a match and to update the list of provider matches synchronously. Emitted when user requests to delete a local history zero-prefix suggestion. tommycli@chromium.org manukh@chromium.org jdonnelly@chromium.org This tracks how unstable omnibox matches are. Typically, each input change is followed by a single synchronous update, and then approximately 3 asynchronous updates. There may be more or less asynchronous updates depending on which providers are enabled for the particular platform/user, the input (e.g. some providers don't run for 0-length inputs), and how quickly the input changes occur (e.g. if the inputs update rapidly, slower async providers won't have time to provide updates before they're aborted by newer inputs). For each autocomplete update, this histogram logs whether any match position changes. If 1 or more matches changed, logs true once; otherwise false once. When the number of matches changes, matches that are removed are considered a 'change', while matches that are appended are less disruptive to the user and not considered a 'change'. This is because this metric is primarily interested in how many times the user could be looking at a match with the intention of selecting it, and then it surprisingly goes away. Note that merely adding a new provider will increase these counts. That's intentional. Swapping matches out annoys users, and to avoid this, we should do some things like: caching, update coalescing, update deferral until next keystroke, provider timeouts, etc. This metric is designed to be normalized by the True count of Omnibox.Start.WantAsyncMatches, which will yield the number of match changes per keystroke or other user gesture. There're the related Omnibox.MatchStability2.InAnyPosition.* slices for tracking match instability for all, asynchronous, and synchronous updates. This histogram can be considered a boolean analogue to the Omnibox.MatchStability2.Index.* histogram which tracks which matches changed. {OmniboxAutocompleteUpdateSlice} tommycli@chromium.org manukh@chromium.org jdonnelly@chromium.org This tracks how unstable omnibox matches are. Typically, each input change is followed by a single synchronous update, and then approximately 3 asynchronous updates. There may be more or less asynchronous updates depending on which providers are enabled for the particular platform/user, the input (e.g. some providers don't run for 0-length inputs), and how quickly the input changes occur (e.g. if the inputs update rapidly, slower async providers won't have time to provide updates before they're aborted by newer inputs). For each match change, this histogram logs the index of the match change. Match changes occur when any match position changes; i.e. is set to a new match different from the old match in that position, even if the new match was contained in the old matches set. If multiple matches change in a single autocomplete update, logs each. If none change in a single autocomplete update, this histogram is not logged. We primarily care about the default match, which is logged at index 0. All indices matter though, so we track the other indices too. When the number of matches changes, matches that are removed are considered a 'change', while matches that are appended are less disruptive to the user and not considered a 'change'. This is because this metric is primarily interested in how many times the user could be looking at a match with the intention of selecting it, and then it surprisingly goes away. Note that merely adding a new provider will increase these counts. That's intentional. Swapping matches out annoys users, and to avoid this, we should do some things like: caching, update coalescing, update deferral until next keystroke, provider timeouts, etc. This metric is designed to be normalized by the True count of Omnibox.Start.WantAsyncMatches, which will yield the number of match changes per keystroke or other user gesture. There're the related Omnibox.MatchStability2.Index.* slices for tracking match instability for all, asynchronous, and synchronous updates. There's the related Omnibox.MatchStability2.InAnyPosition histogram for tracking whether any match changed per autocomplete update. {OmniboxAutocompleteUpdateSlice} peilinwang@google.com woa-performance-bugs+jank@google.com Records the number of characters visible in the omnibox. Recorded once per call to UrlBar.calculateVisibleHint(). jdonnelly@chromium.org mpearson@chromium.org chrome-omnibox-team@google.com The number of terms in the text the user entered in the omnibox when they used the omnibox to go somewhere. Terms are defined by splitting on whitespace. All values larger than 6 are recorded in bucket 6. In M-74 and earlier, this was only recorded if the user had no incognito windows open. In M-75 and later, it records all omnibox interactions regardless, in order to be consistent with how user actions and histograms are handled. cch@chromium.org suggest-2g@google.com Time it takes for omnibox on device head provider to fulfill the asynchronous suggest request. cch@chromium.org suggest-2g@google.com The number of results returned by the on device head model which matches the user input. Note all numbers greater than 5 are recorded in bucket 5. asvitkine@chromium.org mpearson@chromium.org jdonnelly@chromium.org Records the time to paint the omnibox contents. This is a subcomponent of Omnibox.CharTypedToRepaintLatency. Implemented on desktop platforms. jdonnelly@chromium.org chrome-omnibox-team@google.com The number of paste commands on the text in the omnibox. Reported every time a paste command is done. jdonnelly@chromium.org orinj@chromium.org chrome-omnibox-team@google.com Counts the type (concept ID) of omnibox pedals shown when the user used the omnibox to go somewhere. Pedals are a button attached to a search suggestion that either navigate to a chrome:// URL or trigger a Chrome action (for example, opening the page translation dialog or opening a new window in Incognito mode). Note that, while it is uncommon, it is possible for multiple pedals to be shown at one time so this can't be used as a count of how many times a pedal or pedals were present. This histogram is related to Omnibox.SuggestionUsed.Pedal. jdonnelly@chromium.org mpearson@chromium.org chrome-omnibox-team@google.com The length of time taken by the {OmniboxProvider}Provider's synchronous pass. jdonnelly@chromium.org mpearson@chromium.org chrome-omnibox-team@google.com Time it takes for the omnibox to become responsive to user input after the user has typed N characters (for 'Omnibox.QueryTime2.N') or any length input (for 'Omnibox.QueryTime2'). This measures the time it takes to start all autocomplete providers (but not wait for the asynchronous ones to finish). tommycli@chromium.org chrome-omnibox-team@google.com Emits the position of any "Resume Journey" omnibox actions shown in the popup. The position index is 0-based, so an action shown in the topmost match is logged as value 0, consistent with Omnibox.SelectedPosition. Note that these are emitted when the user uses the omnibox to go somewhere. When that happens, one count is emitted per action visible in the popup. These are NOT emitted per-keystroke, in order to make this a suitable denominator for Omnibox.SuggestionUsed.ResumeJourney. junzou@chromium.org chrome-intelligence-core@google.com Emits the score of matched cluster keyword of any "Resume Journey" omnibox actions shown in the popup. Note that these are emitted when the user uses the omnibox to go somewhere. When that happens, one keyword type is emitted per action visible in the popup. These are NOT emitted per-keystroke, in order to make this a suitable denominator for Omnibox.SuggestionUsed.ResumeJourney.ClusterKeywordScore. junzou@chromium.org chrome-intelligence-core@google.com Emits the type of the matched cluster keyword of any "Resume Journey" omnibox actions shown in the popup. The types indicate the sources from which the keywords are generated in the history clustering backend. Note that these are emitted when the user uses the omnibox to go somewhere. When that happens, one keyword type is emitted per action visible in the popup. These are NOT emitted per-keystroke, in order to make this a suitable denominator for Omnibox.SuggestionUsed.ResumeJourney.ClusterKeywordType. junzou@chromium.org chrome-intelligence-core@google.com Emits the top entity collection of any "Resume Journey" omnibox actions shown in the popup. The collections are high-level verticals the keywords belong to. Note that these are emitted when the user uses the omnibox to go somewhere. When that happens, one collection is emitted per action visible in the popup. These are NOT emitted per-keystroke, in order to make this a suitable denominator for Omnibox.SuggestionUsed.ResumeJourney.PageEntityCollection. manukh@chromium.org jdonnelly@chromium.org chrome-omnibox-team@google.com Measures the number of sessions for which each type of rich autocompletion triggered; does not measure the number of times each type of rich autocompletion triggered. As the user types into the omnibox, the omnibox suggestions are updated. Some of those suggestions may have rich autocompletion. No suggestion may have multiple types of rich autocompletion. A type of rich autocompletion is considered triggered only if the default suggestion has rich autocompletion. If a non-default suggestion has rich autocompletion, does not suffice. If any input during the session triggered a rich autocompletion type, that type is considered triggered for the entire session. Recorded multiple times per omnibox session (i.e. per user selecting a suggestion). Recorded once for each type of rich autocompletion that triggered during the session (i.e. while the user was typing their input). If a rich autocompletion type triggered multiple times during the session (very common), it will be logged exactly once. Hence, it's impossible for any single bucket's count to be larger than the count of omnibox sessions or Omnibox.RichAutocompletion.Triggered.Any. If multiple rich autocompletion types triggered during the session (less common), each will be logged. Hence, it's possible for the total count of this histogram to be larger than the count of omnibox sessions. manukh@chromium.org jdonnelly@chromium.org chrome-omnibox-team@google.com Measures the number of sessions for which any type of rich autocompletion triggered. As the user types into the omnibox, the omnibox suggestions are updated. Some of those suggestions may have rich autocompletion. No suggestion may have multiple types of rich autocompletion. A type of rich autocompletion is considered triggered only if the default suggestion has rich autocompletion. If a non-default suggestion has rich autocompletion, does not suffice. If any input during the session triggered a rich autocompletion type, that type is considered triggered for the entire session. The related "Omnibox.RichAutocompletion.Triggered" histogram breaks down the trigger count for each rich autocompletion type. Recorded exactly once per omnibox session (i.e. per user selecting a suggestion). peilinwang@google.com woa-performance-bugs+jank@google.com Measures the amount of time it takes to call UrlBar.scrollToTLD(). Logged once per call. ender@google.com chrome-search@google.com Whenever Ctrl+Enter is used to complete an interaction with the Omnibox, this metric captures whether the keystroke was used to initiate navigation to a website (by affixing ".com") or execute a search query (if affixing ".com" would not produce a valid URL). Recorded each time Ctrl+Enter is used to complete an interaction. ender@google.com chrome-search@google.com Whether Ctrl+Enter key combination was used to complete an interaction with the Omnibox. Recorded each time user accepts Omnibox text via Keyboard. jdonnelly@chromium.org roagarwal@chromium.org chrome-incognito@google.com This histogram records the number of searches done from omnibox using default search engine on desktop and android devices, sliced based on being in regular mode or private modes. mahmadi@chromium.org chrome-omnibox-team@google.com Records the length of the gs_lcrp= param in the Search Result Page request URLs. Recorded every time a Search Result Page URL is generated in the omnibox/realbox. This happens either when the user selects a suggestion in order to load Search Result Page or when a proactive request is made to prefetch the Search Result Page. jdonnelly@chromium.org mpearson@chromium.org chrome-omnibox-team@google.com The type of search engine associated with a match opened from the omnibox. {Population} ryansturm@chromium.org chrome-omnibox-team@google.com Records whether a "search what you type" omnibox navigation was also a {HistoryOrSuggest} type. Recorded when the user navigates to a search via typing the full query into the omnibox. lingqi@chromium.org ryansturm@chromium.org chrome-prerendering@google.com Recorded when posting a task to destroy StreamingSearchPrefetchURLLoader. Used to track whether the native forwarding works as expected. Together with Omnibox.SearchPreload.ForwardingResult.WasServedToPrerender, these two metrics are used to monitor whether shareable cache does not break the forwarding logic. lingqi@chromium.org ryansturm@chromium.org chrome-prerendering@google.com Recorded when posting a task to destroy StreamingSearchPrefetchURLLoader. Together with Omnibox.SearchPreload.ForwardingResult.NotServedToPrerender, these two metrics are used to monitor whether shareable cache does not break the forwarding logic. lingqi@chromium.org ryansturm@chromium.org chrome-prerendering@google.com Tracks whether the StreamingSearchPrefetchURLLoader::ResponseReader can successfully serving to prerender clients. Recorded when a `StreamingSearchPrefetchURLLoader::ResponseReader` is destroyed. Obsolete as of M112. jdonnelly@chromium.org mpearson@chromium.org chrome-omnibox-team@google.com The number of matches returned by SearchProvider. Emitted on every call to SearchProvider::Start(), which effectively means every key stroke in the omnibox. peilinwang@google.com woa-performance-bugs+jank@google.com Measures the amount of time it takes to call UrlBar.setText(). Logged once per call. peilinwang@google.com woa-performance-bugs+jank@google.com Records the length of the text to set when calling UrlBar.setText(). Recorded once per call to setText(). peilinwang@google.com woa-performance-bugs+jank@google.com Records whether the truncated text was enough to fill the urlbar. Recorded every time urlbar text is updated and truncation was applied. manukh@chromium.org chrome-omnibox-team@google.com For each autocomplete input, the shortcut provider 1) searches its index for all matching shortcuts, 2) dedupes them, 3) prunes them to 3, 4) and lastly converts them to `AutocompleteMatch`s. This metric records the number of duplicates per shortcut found in step (2). See the related `Omnibox.Shortcuts.NumberOfUniqueShortcutsIterated`. Recorded once per unique shortcut found each time the shortcut provider is ran with want-async input - i.e. the user editing the omnibox text. But some inputs (e.g. starter pack scopes) won't run the shortcut provider, and some actions (e.g. typing in chrome://omnibox) will also run the shortcut provider. manukh@chromium.org chrome-omnibox-team@google.com For each autocomplete input, the shortcut provider 1) searches its index for all matching shortcuts, 2) dedupes them, 3) prunes them to 3, 4) and lastly converts them to `AutocompleteMatch`s. This metric records the number of unique matching shortcuts (i.e. after step 2, before step 3). See the related `Omnibox.Shortcuts.NumberOfDuplicatesPerShortcutIterated`. Recorded once each time the shortcut provider is ran with want-async input - i.e. the user editing the omnibox text. But some inputs (e.g. starter pack scopes) won't run the shortcut provider, and some actions (e.g. typing in chrome://omnibox) will also run the shortcut provider. tommycli@chromium.org jdonnelly@chromium.org Whether asynchronous matches are requested. Recorded every time AutocompleteController::Start() is called. False counts roughly correspond to non-interactive invocations of AutocompleteController - like when we are classifying user text as a Search or URL. True counts roughly correspond to keystrokes or on-focus events, and can be used as a normalizer for per-user-gesture metrics. jdonnelly@chromium.org mpearson@chromium.org chrome-omnibox-team@google.com Counts how often omnibox suggestions are used, and in the case an Answer suggestion was selected - records the type of an answer. This histogram is useful because there is no distinct suggestion type recorded for answers so their use is not recorded by Omnibox.SuggestionUsed.ProviderAndResultType. nyquist@chromium.org ssid@chromium.org tommycli@chromium.org chrome-segmentation-platform@google.com For each user navigation using the omnibox, logs the selected omnibox result as one of a limited set of result types, based on the autocomplete match. As an example, HISTORY_URL and HISTORY_TITLE are both summarized into the Url summarized result type. This histogram intended to match the server-side histograms that use OmniboxSummarizedResultType such as Omnibox.SuggestionUsed.SearchVsUrl. This is also calculated on the client side to enable functionality that is based on listening to local metrics at runtime. Logged whenever a user navigates using the omnibox. gangwu@chromium.org mpearson@chromium.org jdonnelly@chromium.org Measures whether the suggestion that was selected by the user offered a tab match. A tab match occurs when the URL specified by the suggestion matches the URL navigated to by another tab in the browser, within the margin of our URL stripping scheme. This metric typically only makes sense with respect to Omnibox.SuggestionUsed.OfferedTabMatch. Only recent versions of the browser send the metric. Therefore, other metrics may include datapoints not covered by this metric (unless one narrows the data by version.) In M-74 and earlier, this was only recorded if the user had no incognito windows open. In M-75 and later, it records all omnibox interactions regardless, in order to be consistent with how user actions and histograms are handled. jdonnelly@chromium.org orinj@chromium.org chrome-omnibox-team@google.com Counts how often omnibox pedals are used and records the type (concept ID) of the pedal used. Pedals are a button attached to a search suggestion that either navigate to a chrome:// URL or trigger a Chrome action (for example, opening the page translation dialog or opening a new window in Incognito mode). This histogram is related to Omnibox.PedalShown. tommycli@chromium.org chrome-omnibox-team@google.com chrome-journeys@google.com Logs the position of "Resume Journey" omnibox actions actually selected by the user. The position index is 0-based, so a selected action in the topmost match is logged as value 0, consistent with Omnibox.SelectedPosition. This can be combined with Omnibox.ResumeJourneyShown to compute the Usage (or CTR) of the "Resume Journey" omnibox action on a per-position basis. It's safe to compare the total count of this metric to the total count of Omnibox.ResumeJourneyShown because within a single popup, only zero or one "Resume Journey" can be shown. If multiple actions start being shown in the popup, comparing to the total count becomes misleading to interpret as a CTR. sophiechang@chromium.org chrome-journeys@google.com Logs the score of the matched cluster keyword of "Resume Journey" omnibox actions actually selected by the user. This can be combined with Omnibox.ResumeJourneyShown.ClusterKeywordScore to compute the precision (or CTR) of the "Resume Journey" omnibox action on a keyword score bucket. It's safe to compare the total count of this metric to the total count of Omnibox.ResumeJourneyShown.ClusterKeywordScore because within a single popup, only zero or one "Resume Journey" can be shown. If multiple actions start being shown in the popup, comparing to the total count becomes misleading to interpret as a CTR. sophiechang@chromium.org chrome-journeys@google.com Logs the type of the matched cluster keyword of "Resume Journey" omnibox actions actually selected by the user. This can be combined with Omnibox.ResumeJourneyShown.ClusterKeywordType to compute the precision (or CTR) of the "Resume Journey" omnibox action on a keyword type basis. It's safe to compare the total count of this metric to the total count of Omnibox.ResumeJourneyShown.ClusterKeywordType because within a single popup, only zero or one "Resume Journey" can be shown. If multiple actions start being shown in the popup, comparing to the total count becomes misleading to interpret as a CTR. sophiechang@chromium.org chrome-journeys@google.com Emitted when the user uses the omnibox to go somewhere AND the "Resume Journey" chip triggered by {ClusterKeywordTypeLabel} was offered in any position. This metric records "Selected" if the offered action chip was selected, and records "Not Selected" if the offered action chip was not selected. Therefore the "Selected" bucket proportion is the CTR of the action chip. Higher is better. The count of this bucket matches the {ClusterKeywordTypeLabel} bucket of Omnibox.SuggestionUsed.ResumeJourney.ClusterKeywordType. The total count of both buckets is how many times we've offered the chip triggered by {ClusterKeywordTypeLabel} in total, and matches the {ClusterKeywordTypeLabel} bucket of Omnibox.ResumeJourneyShown.ClusterKeywordType. sophiechang@chromium.org chrome-journeys@google.com Logs the top entity collection of "Resume Journey" omnibox actions actually selected by the user. This can be combined with Omnibox.ResumeJourneyShown.PageEntityCollection to compute the precision (or CTR) of the "Resume Journey" omnibox action on a per-collection basis. It's safe to compare the total count of this metric to the total count of Omnibox.ResumeJourneyShown.PageEntityCollection because within a single popup, only zero or one "Resume Journey" can be shown. If multiple actions start being shown in the popup, comparing to the total count becomes misleading to interpret as a CTR. sophiechang@chromium.org chrome-journeys@google.com Emitted when the user uses the omnibox to go somewhere AND the "Resume Journey" chip triggered by {CollectionLabel} was offered in any position. This metric records "Selected" if the offered action chip was selected, and records "Not Selected" if the offered action chip was not selected. Therefore the "Selected" bucket proportion is the CTR of the action chip. Higher is better. The count of this bucket matches the {CollectionLabel} bucket of Omnibox.SuggestionUsed.ResumeJourney.PageEntityCollection. The total count of both buckets is how many times we've offered the chip triggered by {CollectionLabel} in total, and matches the {CollectionLabel} bucket of Omnibox.ResumeJourneyShown.PageEntityCollection. tommycli@chromium.org chrome-journeys@google.com Emitted when the user uses the omnibox to go somewhere AND the "Resume Journey" chip was offered in any position. This metric records "Selected" if the offered action chip was selected, and records "Not Selected" if the offered action chip was not selected. Therefore the "Selected" bucket proportion is the CTR of the action chip. Higher is better. The count of this bucket matches the sum of Omnibox.SuggestionUsed.ResumeJourney. The total count of both buckets is how many times we've offered the chip total, and matches the sum of Omnibox.ResumeJourneyShown. manukh@chromium.org jdonnelly@chromium.org chrome-omnibox-team@google.com Measures the number of sessions for which each type of rich autocompletion was used. As the user types into the omnibox, the omnibox suggestions are updated. Some of those suggestions may have rich autocompletion. No suggestion may have multiple types of rich autocompletion. A type of rich autocompletion is considered triggered only if the default suggestion has rich autocompletion. If a non-default suggestion has rich autocompletion, does not suffice. If any input during the session triggered a rich autocompletion type, that type is considered triggered for the entire session. Recorded exactly once per omnibox session (i.e. per user selecting a suggestion). jdonnelly@chromium.org mpearson@chromium.org chrome-omnibox-team@google.com Measures the time from navigation start to first meaningful paint. Only recorded for a search query suggestion selected from the omnibox. spelchat@chromium.org chrome-brapp-loading@google.com chrome-omnibox-team@google.com Measures the time from the user selecting the omnibox suggestion and the navigation start time getting ticked. Only recorded for a search query suggestion selected from the omnibox. jdonnelly@chromium.org mpearson@chromium.org chrome-omnibox-team@google.com Measures the time from navigation start to first contentful paint. Only recorded for a search query suggestion selected from the omnibox. lingqi@chromium.org chrome-prerendering@google.com chrome-omnibox-team@google.com Measures the time from navigation start to largest contentful paint (https://github.com/w3c/largest-contentful-paint). Only recorded for a search query suggestion selected from the omnibox. spelchat@chromium.org chrome-brapp-loading@google.com chrome-omnibox-team@google.com Measures the time from navigation to the time the first largest contentful paint candidate (https://github.com/w3c/largest-contentful-paint#the-last-candidate) above 2 seconds. Only recorded for a search query suggestion selected from the omnibox. Note that this histogram is most useful for use with Chrometto in order to record a high LCP trace as soon as it happens (rather than only once the slow page loads gets discarded). gangwu@chromium.org mpearson@chromium.org jdonnelly@chromium.org Measures whether the suggestion that was selected by the user offered a tab match and specifically that the user chose the tab match over normal navigation. A tab match occurs when the URL specified by the suggestion matches the URL navigated to by another tab in the browser, within the margin of our URL stripping scheme. By selecting the tab match, the browser simply switches focus to the matching tab. This metric typically only makes sense with respect to Omnibox.SuggestionUsed.OfferedTabMatch. Only recent versions of the browser send the metric. Therefore, other metrics may include datapoints not covered by this metric (unless one narrows the data by version.) This metric should be stricly less than Omnibox.SuggestionUsed.OfferedTabMatch, and is typically used to discover the percentage of offered tab matches that the user accepts. In M-74 and earlier, this was only recorded if the user had no incognito windows open. In M-75 and later, it records all omnibox interactions regardless, in order to be consistent with how user actions and histograms are handled. jdonnelly@chromium.org mpearson@chromium.org chrome-omnibox-team@google.com Measures the time from navigation start to first meaningful paint. Only recorded for a URL suggestion selected from the omnibox. spelchat@chromium.org chrome-brapp-loading@google.com chrome-omnibox-team@google.com Measures the time from the user selecting the omnibox suggestion and the navigation start time getting ticked. Only recorded for a URL suggestion selected from the omnibox. jdonnelly@chromium.org mpearson@chromium.org chrome-omnibox-team@google.com Measures the time from navigation start to first contentful paint. Only recorded for a URL suggestion selected from the omnibox. robertlin@chromium.org chrome-prerendering@google.com chrome-omnibox-team@google.com Measures the time from navigation start to largest contentful paint. Only recorded for a URL suggestion selected from the omnibox. mpearson@chromium.org jdonnelly@chromium.org cch@chromium.org The time elapsed between the sending of a suggest request to Google until the time the request was returned with status==failed. Ignores requests that were canceled before being returned. mpearson@chromium.org jdonnelly@chromium.org cch@chromium.org The time elapsed between the sending of a suggest request to Google until the time the request was returned with status==success. Ignores requests that were canceled before being returned. 3/2022; feature launched. manukh@chromium.org jdonnelly@chromium.org chrome-omnibox-team@google.com The number of images prefetched by the search provider on retrieving a successful search response. mpearson@chromium.org jdonnelly@chromium.org cch@chromium.org Counts about the number of suggest requests the omnibox sent, invalidated, and replies received. ender@google.com mahmadi@chromium.org The index of the item in the MostVisited carousel that the user deleted. Zero-based: the first item in the carousel is recorded in bucket 0, second in bucket 1 and so on. Items beyond 16 will be stored in the overflow bucket (15). Recorded every time the user initiates deletion of a suggest tile, even if the operation is later canceled. ender@google.com mahmadi@chromium.org Records the type of the SuggestTile that the user deleted. Most Visited URLs record a URL, and Organic Repeatable Queries record a Search type. Recorded every time the user initiates deletion of a suggest tile, that results with tile removal. Not recorded if the user chooses to cancel the action. ender@google.com mahmadi@chromium.org The index of the item in the MostVisited carousel that the user opened. Zero-based: the first item in the carousel is recorded in bucket 0, second in bucket 1 and so on. Items beyond 16 will be stored in the overflow bucket (15). Recorded every time the user selects the tile to initiate navigation, even if the navigation could not be completed (eg. target site did not load). ender@google.com mahmadi@chromium.org Records the type of the SuggestTile that the user interacted with. Most Visited URLs record a URL, and Organic Repeatable Queries record a Search type. Recorded every time the user selects the tile to initiate navigation, even if the navigation could not be completed (eg. target site did not load). Note: possible recording gap between 2023/06 and 2023/08. ender@google.com mahmadi@chromium.org Total count of offered {SuggestTileType} tile types. Recorded every time the current context is eligible to receive MostVisitedTiles, even if there were no tiles of any kind to show. Note: possible recording gap between 2023/06 and 2023/08. gangwu@chromium.org jdonnelly@chromium.org mpearson@chromium.org chrome-omnibox-team@google.com This histogram reports the amount of time it takes to match one set of suggestions against all open tabs. This metric is collected once for every time a new set of suggestions is received. This includes the Omnibox Focus event (when zero-prefix suggestions are offered) and every subsequent key stroke, including deletion, backspace etc. Each keystroke produces multiple events offering autocomplete matches: a) one synchronous, when Start() is called. b) any number of intermediate events, depending on whether Providers request the update. c) one at the end, when the process completes. Since these are produced about 3 times per keystroke, the total time consumed per Omnibox interaction is - optimistically - 3*Omnibox.TypedLength*TabMatchTime. this excludes all the time spent formulating the query, such as correcting the input. This metric does not express the cumulative time spent matching the tabs (because we do not have a metric that counts the number of suggestions the user actually fetched - eg. we do not count the number of suggestions received when user pressed the backspace key). This histogram only records metrics on machines with high-resolution clocks. yusuyoutube@google.com benwgold@google.com fgorski@chromium.org wylieb@chromium.org lens-chrome@google.com Logs the elapsed time between the time when user enters the Google Lens and the time when user dismisses Lens and returns to Chrome. Only logged on Android. Replaced with Omnibox.GroupId.ToggledOff in M114. mahmadi@google.com chrome-omnibox-team@google.com Counts the number of times a suggestion group ID is prevented form appearing in the results. This is recorded when the user hides a set of grouped suggestions, such as proactive zero-suggest, in the omnibox UI using the control in the grouped suggestions' header. Replaced with Omnibox.GroupId.ToggledOn in M114. mahmadi@google.com chrome-omnibox-team@google.com Counts the number of times a suggestion group ID is permitted to appear in the results. This is recorded when the user makes a set of previously hidden grouped suggestions, such as proactive zero-suggest, visible in the omnibox UI using the control in the grouped suggestions' header. mpearson@chromium.org jdonnelly@chromium.org chrome-omnibox-team@google.com The scheme of the destination URL for the selected omnibox suggestion. This could be a what-you-typed suggestion (if the user fully typed a URL), an inline autocompletion, or something in the dropdown, anything as long as it's a URL navigation, not a search query. Excludes omnibox URL navigations that are effectively reloads as well as cut-and-pastes of URLs. Also does not take into account any redirects that destination URL may do before finally serving a page. In other words, it's the scheme of the first request in a possible chain. Due to a bug, recorded paste-and-go actions through M-88 (inclusive). Excluded in M-89+. mahmadi@chromium.org yoangela@chromium.org chrome-desktop-search@google.com The amount of time it takes to execute the URL scoring model for all the eligible matches. Emitted when all the providers are done and there is at least one eligible match for model execution. mahmadi@chromium.org yoangela@chromium.org chrome-desktop-search@google.com The number of matches the URL scoring model was executed for. Emitted when all the providers are done and there is at least one eligible match for model execution. tluk@chromium.org chrome-desktop-search@google.com Records the number of milliseconds between the views omnibox popup construction and the first time content for the popup is displayed on the screen. This is emitted once each time the omnibox is opened. It is recorded again if the omnibox has closed and later re-opened. It is not emitted more than once for a single opened instance of the ombibox popup. orinj@chromium.org chrome-desktop-search@google.com Records the number of milliseconds between the webui omnibox popup construction and its first UpdatePopupAppearance call. This measures how long the popup existed before first use. Since the webui begins the asynchronous process of initialization and remote binding immediately on construction, this metric can be seen as a time budget for handler preparation. See also Omnibox.WebUI.HandlerReadyOnFirstAccess. orinj@chromium.org chrome-desktop-search@google.com Records the ready state of the webui handler when first accessed. When this is true, the page loaded and the handler was ready on time. When this is false, it indicates that OmniboxPopupPresenter::GetHandler returned nullptr at least once because the page wasn't loaded before the first attempt to communicate via the handler. Recorded once per omnibox popup presenter instantiation, when the webui omnibox popup page handler is first used; so, at most once per browser window. pnoland@google.com ender@google.com chrome-omnibox-team@google.com Counts the type (concept ID) of omnibox zero-prefix on-focus type requests. The omnibox of Chrome on Android measures use of different focus types while retrieving zero prefix suggestions on supported platforms, and records exactly once per every zero prefix suggestions request. pnoland@chromium.org chrome-omnibox-team@google.com The number of results (either query or URL) from ZeroSuggest. This is emitted every time ZeroSuggestProvider produces results on omnibox focus or on omnibox input clobber, either from the cache or from the remote response. Obsolete as of M106. mpearson@chromium.org jdonnelly@chromium.org Whether the user has settings configured so that the current page URL can be sent to the suggest server to request contextual suggestions. For example, this is only supported for users who have Google as their default search engine (unmodified version of Google), have search suggest enabled, are signed-in and syncing without a custom passphrase, and don't have an incognito window open. There are other criteria too. Recorded on focus in the omnibox if there is default search provider and we've constructed a suggest URL. Some additional guidelines: if an incognito window is open, all focus events will go into the "generally ineligible" bucket. Likewise, if the current page is a search results page, we don't allow contextual suggestions either so focus events on those pages go in the "generally ineligible" bucket. The difference between "eligible" and "generally eligible but not this time" depends only the properties of the current URL. Recorded regardless of whether contextual or non-contextual zero suggest is currently enabled on the user's platform. However, if zero suggest (in all forms) is entirely disabled, the user will be perpetually ineligible. This histogram captures all scenarios that are generally ineligible for zero suggest suggestions (including incognito mode), and reports eligibility for suggestions even if kOmniboxPopupShortcutIconsInZeroState is enabled (meaning: suggestions are not shown, but current URL is eligible). pnoland@chromium.org chrome-omnibox-team@google.com The number of query results returned from ZeroSuggest. This is emitted every time ZeroSuggestProvider produces results on omnibox focus or on omnibox input clobber, either from the cache or from the remote response. pnoland@chromium.org chrome-omnibox-team@google.com The number of URL results returned from ZeroSuggest. This is emitted every time ZeroSuggestProvider produces results on omnibox focus or on omnibox input clobber, either from the cache or from the remote response. mahmadi@chromium.org chrome-omnibox-team@google.com Emitted in ZeroSuggestCacheService, records the estimated dynamically allocated memory usage of the ZeroSuggestProvider cache every time the prefetch or the non-prefetch zero-suggest response is stored in the cache on eligible page load / page switch events and when the user has focused or cleared the omnibox but has not typed in yet, respectively. Note that the max number of responses stored in the cache is always capped. The cache size can build up as more responses are stored. The cache size can also decrease as responses are removed from the cache when suggestions are deleted from the omnibox or when the user clears the browsing data. Hence, due to the dynamic nature of the cache size, this histogram is recorded on every store event as opposed to the cache destructor. mahmadi@chromium.org chrome-omnibox-team@google.com Logs whether ZeroSuggestProvider is eligible to request and display zero- prefix suggestions, and if not, why not. Emitted for both omit_asynchronous_matches_ true (i.e, Autocomplete classifier) and false inputs (i.e, when the user has typed in, focused the omnibox, or cleared it but has not typed in yet) in ZeroSuggestProvider::Start(). Also emitted on applicable page load/switch events in ZeroSuggestProvider::StartPrefetch(). mahmadi@chromium.org chrome-omnibox-team@google.com Enumerates the events in the ZeroSuggestProvider as a result of {RequestType} for {ResultType} (i.e., suggestions shown when the user has focused or cleared the omnibox but has not typed in yet). Obsolete as of M105. mahmadi@chromium.org chrome-omnibox-team@google.com Counts the number of non-prefetch zero suggest requests (requests for suggestions when the user has focused but not modified the omnibox) the omnibox sent, requests that were invalidated, responses that were received, the subset of responses that were loaded from HTTP cache, and finally the subset of responses that ended up updating the displayed results. Obsolete as of M105. mahmadi@chromium.org chrome-omnibox-team@google.com Records the time elapsed between non-prefetch zero suggest requests (requests for suggestions when the user has focused but not modified the omnibox) sent and the responses received; whether or not the response is loaded from HTTP cache. Obsolete as of M105. mahmadi@chromium.org chrome-omnibox-team@google.com Counts the number of prefetch zero suggest requests (requests for suggestions when the user opens a new tab or brings one to foreground) the omnibox sent, requests that were invalidated, responses that were received, the subset of responses that were loaded from HTTP cache, and finally the subset of responses that ended up updating the displayed results. Obsolete as of M105. mahmadi@chromium.org chrome-omnibox-team@google.com Records the time elapsed between prefetch zero suggest requests (requests for suggestions when the user opens a new tab or brings one to foreground) sent and the responses received; whether or not the response is loaded from HTTP cache. spelchat@chromium.org chrome-brapp-loading@google.com chrome-omnibox-team@google.com {SearchPrefetch} Measures the time between the user clicks the Omnibox suggestion and the moment the navigation is intercepted by the Search Prefetch code. {NavigationPrefetchOnly} ryansturm@chromium.org chrome-omnibox-team@google.com {SearchPrefetch} Whether a search prefetch request finished with a success status or a status that could be served. Recorded when prefetch requests receive headers. {SuggestionPrefetchOnly} spelchat@chromium.org chrome-brapp-loading@google.com chrome-omnibox-team@google.com {SearchPrefetch} Measures the time between the navigation is intercepted by the Search Prefetch code and the moment the Search Prefetch URL loader has finished forwarding the whole response to the navigation URL loader. This time includes finishing fetching the Search response and any queuing delays in the UI thread preventing the Search Prefetch URL loader from forwarding bytes to the navigation code. ryansturm@chromium.org chrome-omnibox-team@google.com {SearchPrefetch} Whether the prefetch URL and the navigation URL are an exact match including query params in the URL. Recorded upon navigation to a matching search term. {NavigationPrefetchOnly} This is deprecated as of 07/2022. Please use Omnibox.SearchPrefetch.PrefetchEligibilityReason2.{NavigationOrSuggestionPrefetch}. ryansturm@chromium.org chrome-omnibox-team@google.com {SearchPrefetch} The reason that a search prefetch request issued when a user navigates to a suggestion was not eligible to be prefetched (or 'Prefetched' if it was prefetched). {NavigationOrSuggestionPrefetch} ryansturm@chromium.org chrome-omnibox-team@google.com {SearchPrefetch} The reason that a search prefetch request issued when a user navigates to a suggestion was not eligible to be prefetched (or 'Prefetched' if it was prefetched). {NavigationOrSuggestionPrefetch} ryansturm@chromium.org chrome-omnibox-team@google.com {SearchPrefetch} The status of a Search Prefetch request when it is deleted by expiration or by serving the request. {NavigationOrSuggestionPrefetch} ryansturm@chromium.org lingqi@chromium.org chrome-omnibox-team@google.com chrome-prerendering@google.com {SearchPrefetch} The reason that a navigation request could or could not serve a search prefetch request. {PrerenderOrNavigation} This is deprecated as of 08/2022. Please use Omnibox.SearchPrefetch.PrefetchServingReason2{PrerenderOrNavigation}. ryansturm@chromium.org lingqi@chromium.org chrome-omnibox-team@google.com chrome-prerendering@google.com {SearchPrefetch} The reason that a navigation request could or could not serve a search prefetch request. {PrerenderOrNavigation} Removed since M112 in favor of Omnibox.{SearchPrefetch}.ReceivedServableResponse2.{FallbackOrInitial}.{NavigationOrSuggestionPrefetch}. ryansturm@chromium.org chrome-omnibox-team@google.com {SearchPrefetch} This histogram records whether an activated (i.e., served) prefetch URL Loader receives a servabled response from the server. {FallbackOrInitial} {NavigationOrSuggestionPrefetch} ryansturm@chromium.org chrome-omnibox-team@google.com chrome-prerendering@google.com {SearchPrefetch} This histogram records whether an activated (i.e., served) prefetch URL Loader receives a servabled response from the server. A servabled reasponse should contain valid headers and body, with any 200 response code. {FallbackOrInitial} {NavigationOrSuggestionPrefetch} ryansturm@chromium.org curranmax@chromium.org chrome-search@google.com {SearchPrefetch} The amount of time it takes to start a prefetch request. {NavigationPrefetchOnly} mahmadi@chromium.org chrome-desktop-search@google.com Records the time taken between a keystroke being typed in the NTP realbox and the time when we're ready to render the results in the DOM. If there are multiple keystrokes before the results are rendered, logs the time since the earliest. This is comparable to Omnibox.CharTypedToRepaintLatency.ToPaint as it does not take into account the time the DOM change task spends in the message loop nor its duration. mahmadi@chromium.org chrome-desktop-search@google.com Records the time taken between AutocompleteController::Observer's OnResultChanged is called and the time when we're ready to render the results in the NTP realbox. If there are multiple calls to OnResultChanged before the results are rendered, logs the time since the earliest. This is comparable to Realbox.CharTypedToRepaintLatency.ToPaint but excludes the time taken to send the input to the browser and process it. Similarly this does not take into account the time the DOM change task spends in the message loop nor its duration.