PerformanceMonitor: Add a new UMA histograms to track average CPU utilization

chrome/browser/performance_monitor/process_metrics_history.cc

Index: chrome/browser/performance_monitor/process_metrics_history.cc
diff --git a/chrome/browser/performance_monitor/process_metrics_history.cc b/chrome/browser/performance_monitor/process_metrics_history.cc
index 9ac206fb36151817a4ccbe446fb28fab2d91d2f6..7bf658deda4821fcdeb4fd72e846bb9e3035ab70 100644
--- a/chrome/browser/performance_monitor/process_metrics_history.cc
+++ b/chrome/browser/performance_monitor/process_metrics_history.cc
@@ -74,9 +74,13 @@ void ProcessMetricsHistory::EndOfCycle() {
 
 void ProcessMetricsHistory::RunPerformanceTriggers() {
   // As an initial step, we only care about browser processes.
-  if (process_type_ != content::PROCESS_TYPE_BROWSER)
+  if (process_type_ != content::PROCESS_TYPE_BROWSER || sample_count_ == 0)
     return;
 
+  UMA_HISTOGRAM_CUSTOM_COUNTS("PerformanceMonitor.AverageCPU.BrowserProcess",
+                              accumulated_cpu_usage_ / sample_count_,
+                              0, 1200, 20);

[Alexei Svitkine (slow), 2013/10/18 21:39:12]

Can you explain why you're using these parameters? What's 1200?

[oystein (OOO til 10th of July), 2013/10/21 14:38:45]

1200 is the maximum value we'd see on a 12-core non-Windows system. It's
reasonably arbitrary, if anyone has any better suggestions.

[Alexei Svitkine (slow), 2013/10/21 14:53:46]

I'd suggest picking a value that's more future-proof (the current maximum of
today is not the maximum of tomorrow).

Alternatively, scale non-Windows values to 100% as the max and use
UMA_HISTOGRAM_PERCENTAGE() (which will also give you distinct buckets for all
values rather than exponential bucket ranges like UMA_HISTOGRAM_CUSTOM_COUNTS
will produce).

[oystein (OOO til 10th of July), 2013/10/21 15:15:05]

Sure, I'll bounce it to 6400 or something. 

[Alexei Svitkine (slow), 2013/10/21 15:27:31]

All right. And you're okay with the exponential bucket ranges provided by this
macro?

+
   // If CPU usage has consistently been above our threshold,
   // we *may* have an issue.
   if (min_cpu_usage_ > kHighCPUUtilizationThreshold)