[PageLoadMetrics] Keep track of Ad Sizes on Pages

chrome/browser/page_load_metrics/observers/ads_page_load_metrics_observer.cc

+// Copyright 2017 The Chromium Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style license that can be
+// found in the LICENSE file.
+
+#include "chrome/browser/page_load_metrics/observers/ads_page_load_metrics_observer.h"
+
+#include <string>
+
+#include "base/feature_list.h"
+#include "base/strings/string_util.h"
+#include "chrome/browser/page_load_metrics/page_load_metrics_util.h"
+#include "content/public/browser/navigation_handle.h"
+#include "content/public/browser/render_frame_host.h"
+#include "content/public/browser/web_contents.h"
+#include "url/gurl.h"
+
+namespace {
+
+const base::Feature kAdsFeature{"AdsMetrics", base::FEATURE_ENABLED_BY_DEFAULT};
+
+bool FrameIsAd(content::NavigationHandle* navigation_handle) {
+  int frame_tree_node_id = navigation_handle->GetFrameTreeNodeId();
+  content::RenderFrameHost* current_frame_host =
+      navigation_handle->GetWebContents()->FindFrameByFrameTreeNodeId(
+          frame_tree_node_id);
+  const std::string& name = current_frame_host->GetFrameName();
+  const GURL& url = navigation_handle->GetURL();
+
+  // Because sub-resource filtering isn't always enabled, and doesn't work
+  // well in monitoring mode (no CSS enforcement), it's difficult to identify
+  // ads. Google ads are prevalent and easy to track, so we'll start by
+  // tracking those. Note that the frame name can be very large, so be careful
+  // to avoid full string searches if possible.
+  // TODO(jkarlin): Track other ad networks that are easy to identify.
+  return base::StartsWith(name, "google_ads_iframe",
+                          base::CompareCase::SENSITIVE) ||
+         base::StartsWith(name, "google_ads_frame",
+                          base::CompareCase::SENSITIVE) ||
+         (url.host_piece() == "tpc.googlesyndication.com" &&
+          base::StartsWith(url.path_piece(), "/safeframe",
+                           base::CompareCase::SENSITIVE));
+}
+
+}  // namespace
+
+AdsPageLoadMetricsObserver::AdsPageLoadMetricsObserver()
+    : enabled_(base::FeatureList::IsEnabled(kAdsFeature)) {}

[Charlie Harrison, 2017/04/13 19:09:57]

Sorry I don't think I was clear. Why do we even need to add this observer if the
feature isn't enabled? Let's just make a static MaybeCreate() method which
returns nullptr or something if the feature isn't enabled.

[jkarlin, 2017/04/14 17:50:20]

Done.

+
+AdsPageLoadMetricsObserver::~AdsPageLoadMetricsObserver() = default;
+
+page_load_metrics::PageLoadMetricsObserver::ObservePolicy
+AdsPageLoadMetricsObserver::OnCommit(
+    content::NavigationHandle* navigation_handle) {
+  DCHECK(ad_frames_data_.empty());
+
+  if (!enabled_)
+    return STOP_OBSERVING;
+
+  // The main frame is never considered an ad.
+  ad_frames_data_[navigation_handle->GetFrameTreeNodeId()] = nullptr;
+  ProcessDelayedResources(navigation_handle->GetFrameTreeNodeId());
+  return CONTINUE_OBSERVING;
+}
+
+page_load_metrics::PageLoadMetricsObserver::ObservePolicy
+AdsPageLoadMetricsObserver::OnCommitSubFrame(
+    content::NavigationHandle* navigation_handle) {
+  FrameTreeNodeId frame_tree_node_id = navigation_handle->GetFrameTreeNodeId();
+  content::RenderFrameHost* parent_frame_host =
+      navigation_handle->GetRenderFrameHost()->GetParent();
+  DCHECK(parent_frame_host);
+
+  bool top_level_frame = !parent_frame_host->GetParent();

[Charlie Harrison, 2017/04/13 19:09:57]

optional: top_level_frame (and the associated histogram) implies main frame to
me. What do you think about top_level_subframe?

[jkarlin, 2017/04/14 17:50:20]

Agreed, done.

+
+  const auto& id_and_data = ad_frames_data_.find(frame_tree_node_id);
+  if (id_and_data != ad_frames_data_.end()) {
+    // An existing subframe is navigating again.
+    if (id_and_data->second) {
+      // The subframe was an ad to begin with, keep tracking it as an ad.
+      ProcessDelayedResources(frame_tree_node_id);
+      return CONTINUE_OBSERVING;
+    }
+    // This frame was previously not an ad, process it as usual. If it had
+    // any child frames that were ads, those will still be recorded.
+  } else if (top_level_frame) {
+    top_level_frame_count_ += 1;
+  }
+
+  // Determine who the parent frame's ad ancestor is.
+  const auto& parent_id_and_data =
+      ad_frames_data_.find(parent_frame_host->GetFrameTreeNodeId());
+  DCHECK(parent_id_and_data != ad_frames_data_.end());
+  AdFrameData* ad_data = parent_id_and_data->second;
+
+  if (!ad_data && FrameIsAd(navigation_handle)) {
+    // This frame is not nested within an ad frame but is itself an ad.
+    ad_frames_data_storage_.push_back(AdFrameData());
+    ad_data = &ad_frames_data_storage_.back();
+  }
+
+  ad_frames_data_[frame_tree_node_id] = ad_data;
+
+  if (top_level_frame && ad_data)
+    top_level_ad_frame_count_ += 1;
+
+  ProcessDelayedResources(frame_tree_node_id);
+  return CONTINUE_OBSERVING;
+}
+
+page_load_metrics::PageLoadMetricsObserver::ObservePolicy
+AdsPageLoadMetricsObserver::FlushMetricsOnAppEnterBackground(
+    const page_load_metrics::PageLoadTiming& timing,
+    const page_load_metrics::PageLoadExtraInfo& extra_info) {
+  // The browser may come back, but there is no guarantee. To be safe, record
+  // what we have now and ignore future changes to this navigation.
+  if (extra_info.did_commit)
+    RecordHistograms();
+
+  return STOP_OBSERVING;
+}
+
+void AdsPageLoadMetricsObserver::OnLoadedResource(
+    const page_load_metrics::ExtraRequestInfo& extra_request_info) {
+  const auto& id_and_data =
+      ad_frames_data_.find(extra_request_info.frame_tree_node_id);
+  if (id_and_data == ad_frames_data_.end()) {
+    // This resouce is for a frame that hasn't yet committed. It must be the
+    // main document for the frame. Hold onto it and once it commits we'll run
+    // it in ProcessDelayedResources.
+    auto it_and_success = delayed_resources_.insert(std::make_pair(

[Charlie Harrison, 2017/04/13 19:09:57]

delayed_resources_ is a little generic. What do you think about
"ongoing_navigation_resources_" or something to that affect?

[jkarlin, 2017/04/14 17:50:20]

Done.

+        extra_request_info.frame_tree_node_id, extra_request_info));
+    DCHECK(it_and_success.second);
+    return;
+  }
+
+  page_bytes_ += extra_request_info.raw_body_bytes;
+  if (!extra_request_info.was_cached)
+    uncached_page_bytes_ += extra_request_info.raw_body_bytes;
+
+  // Determine if the frame (or its ancestor) is an ad, if so attribute the
+  // bytes to the highest ad ancestor.
+  AdFrameData* ancestor_data = id_and_data->second;
+
+  if (ancestor_data) {
+    ancestor_data->frame_bytes += extra_request_info.raw_body_bytes;
+    if (!extra_request_info.was_cached) {
+      ancestor_data->frame_bytes_uncached += extra_request_info.raw_body_bytes;
+    }
+  }
+}
+
+void AdsPageLoadMetricsObserver::OnComplete(
+    const page_load_metrics::PageLoadTiming& timing,
+    const page_load_metrics::PageLoadExtraInfo& info) {
+  RecordHistograms();
+}
+
+void AdsPageLoadMetricsObserver::RecordHistograms() {
+  if (page_bytes_ == 0)
+    return;
+
+  size_t total_ad_frame_bytes = 0;
+  size_t uncached_ad_frame_bytes = 0;
+
+  for (const AdFrameData& ad_frame_data : ad_frames_data_storage_) {
+    total_ad_frame_bytes += ad_frame_data.frame_bytes;
+    uncached_ad_frame_bytes += ad_frame_data.frame_bytes_uncached;
+
+    PAGE_BYTES_HISTOGRAM("PageLoad.Clients.Ads.Google.Bytes.AdFrameBytes",
+                         ad_frame_data.frame_bytes);
+    PAGE_BYTES_HISTOGRAM(
+        "PageLoad.Clients.Ads.Google.Bytes.AdFrameBytesFromNetwork",
+        ad_frame_data.frame_bytes_uncached);
+    if (ad_frame_data.frame_bytes > 0) {
+      UMA_HISTOGRAM_PERCENTAGE(
+          "PageLoad.Clients.Ads.Google.Bytes.PercentAdFrameBytesFromNetwork",
+          ad_frame_data.frame_bytes_uncached * 100 / ad_frame_data.frame_bytes);
+    }
+  }
+
+  // Don't post UMA for pages that don't have ads.
+  if (ad_frames_data_storage_.empty()) {
+    UMA_HISTOGRAM_COUNTS("PageLoad.Clients.Ads.Google.PageHasNoAds", 1);

[Charlie Harrison, 2017/04/13 19:09:57]

Use the Boolean specific histogram macros, they're cheaper.

[jkarlin, 2017/04/14 17:50:20]

Done.

+    return;
+  }
+
+  UMA_HISTOGRAM_COUNTS_1000("PageLoad.Clients.Ads.Google.TopLevelFrameCount",
+                            top_level_frame_count_);
+  UMA_HISTOGRAM_COUNTS_1000("PageLoad.Clients.Ads.Google.TopLevelAdFrameCount",
+                            top_level_ad_frame_count_);
+
+  DCHECK_LT(0, top_level_frame_count_);  // Because ad frames isn't empty.
+  UMA_HISTOGRAM_PERCENTAGE(
+      "PageLoad.Clients.Ads.Google.PercentTopLevelFramesAreAdFrames",
+      top_level_ad_frame_count_ * 100 / top_level_frame_count_);
+
+  PAGE_BYTES_HISTOGRAM("PageLoad.Clients.Ads.Google.Bytes.AllAdFramesBytes",
+                       total_ad_frame_bytes);
+
+  PAGE_BYTES_HISTOGRAM(
+      "PageLoad.Clients.Ads.Google.Bytes.PageBytesSansAllAdFrames",
+      page_bytes_ - total_ad_frame_bytes);
+
+  PAGE_BYTES_HISTOGRAM("PageLoad.Clients.Ads.Google.Bytes.PageBytes",
+                       page_bytes_);
+  PAGE_BYTES_HISTOGRAM("PageLoad.Clients.Ads.Google.Bytes.PageBytesFromNetwork",
+                       uncached_page_bytes_);
+  if (page_bytes_) {
+    UMA_HISTOGRAM_PERCENTAGE(
+        "PageLoad.Clients.Ads.Google.Bytes.PercentPageBytesFromAllAdFrames",
+        total_ad_frame_bytes * 100 / page_bytes_);
+  }
+
+  UMA_HISTOGRAM_COUNTS_1000("PageLoad.Clients.Ads.Google.AdFrameCount",
+                            ad_frames_data_storage_.size());
+
+  PAGE_BYTES_HISTOGRAM(
+      "PageLoad.Clients.Ads.Google.Bytes.AllAdFramesBytesFromNetwork",
+      uncached_ad_frame_bytes);
+
+  if (total_ad_frame_bytes) {
+    UMA_HISTOGRAM_PERCENTAGE(
+        "PageLoad.Clients.Ads.Google.Bytes.PercentAllAdFramesBytesFromNetwork",
+        uncached_ad_frame_bytes * 100 / total_ad_frame_bytes);
+  }
+  if (uncached_page_bytes_ > 0) {
+    UMA_HISTOGRAM_PERCENTAGE(
+        "PageLoad.Clients.Ads.Google.Bytes."
+        "PercentPageNetworkBytesFromAllAdFrames",
+        uncached_ad_frame_bytes * 100 / uncached_page_bytes_);
+  }
+}
+
+void AdsPageLoadMetricsObserver::ProcessDelayedResources(
+    FrameTreeNodeId frame_tree_node_id) {
+  const auto& frame_id_and_request =
+      delayed_resources_.find(frame_tree_node_id);
+  if (frame_id_and_request == delayed_resources_.end())
+    return;
+  OnLoadedResource(frame_id_and_request->second);
+  delayed_resources_.erase(frame_id_and_request);
+}