[Telemetry] Add an example for adding an end to end Telemetry benchmark

tools/telemetry/examples/benchmarks/v8_metric.py

+# Copyright 2015 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.
+
+from telemetry.value import improvement_direction
+from telemetry.value import scalar
+from telemetry.web_perf.metrics import timeline_based_metric
+
+def _IsMessageLoopEvent(event):
+  return event.name.startswith('v8')
+
+
+class _AverageMessageLoopLatency(scalar.ScalarValue):
+  def __init__(self, value, page, tir_label, none_value_reason=None):
+    super(_AverageMessageLoopLatency, self).__init__(
+      page=page, name='avg_v8_events_latency', value=value,
+      tir_label=tir_label,
+      units='ms', improvement_direction=improvement_direction.DOWN,
+      description=('Average wall-time latency of message loop events during '
+                   'any of the interaction records\' time ranges'),
+      none_value_reason=none_value_reason)
+
+
+class MessageLoopLatencyMetric(timeline_based_metric.TimelineBasedMetric):
+
+  def AddResults(self, model, renderer_thread, interactions, results):
+    v8_events = []
+    for event in model.IterAllEvents(event_predicate=_IsMessageLoopEvent):
+      if timeline_based_metric.IsEventInInteractions(event, interactions):
+        v8_events.append(event)
+
+    if v8_events:
+      avg = (
+          sum(e.duration for e in v8_events)/len(v8_events))
+      results.AddValue(_AverageMessageLoopLatency(
+          value=avg, page=results.current_page,
+          tir_label=interactions[0].label))
+    else:
+      results.AddValue(_AverageMessageLoopLatency(
+          None, page=results.current_page,
+          tir_label=interactions[0].label,
+          none_value_reason='No v8 events found.'))